This action might not be possible to undo. Are you sure you want to continue?
June 2008 Edition | © 2008 WebTrends Inc.
This document and the software described in this document are furnished under and are subject to the terms of a license agreement or a non-disclosure agreement. EXCEPT AS EXPRESSLY SET FORTH IN A LICENSE AGREEMENT, WEBTRENDS INC. PROVIDES THIS DOCUMENT AND THE SOFTWARE DESCRIBED IN THIS DOCUMENT "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. You agree that you shall not loan, sell, or otherwise transfer this document or the software described in this document. Except as expressly set forth in a license agreement, you agree that you shall not reproduce, store in a retrieval system, or transmit in any form or by any means, electronic, mechanical, or otherwise, all or any part of this document or the software described in this document. Some companies, names, and data in this document are used for illustration purposes and do not represent real companies, individuals, or data. This document may include technical inaccuracies or typographical errors. WebTrends Inc. may make improvements in or changes to the software described in this document at any time. © 1996-2008 WebTrends Inc. All rights reserved. U.S. Government Restricted Rights: The software is "commercial software." If the software and documentation are being acquired by or on behalf of the U.S. Government or by a U.S. Government prime contractor or subcontractor (at any tier), in accordance with 48 C.F.R. 227.7202-4 (for Department of Defense (DOD) acquisitions) and 48 C.F.R. 2.101 and 12.212 (for non-DOD acquisitions), the government's rights in the software and documentation, including its rights to use, modify, reproduce, release, perform, display or disclose the software or documentation, will be subject in all respects to the commercial license rights and restrictions provided in the license agreement.
WebTrends Visitor Intelligence™, WebTrends Visitor Intelligence logo™, WebTrends Score™, WebTrends Score logo™, WebTrends Marketing Lab™, WebTrends ml2™, WebTrends ml2 logo™, WebTrends Analytics™, WebTrends Analytics logo™, WebTrends Dymanic Search™, WebTrends Dymanic Search logo™, WebTrends Marketing Warehouse™, WebTrends Explore™, WebPosition®, True Optimization. Real Results.™, True Optimization™, WebTrends Open Exchange™, WebTrends® and the WebTrends logo™ are trademarks or registered trademarks of WebTrends Inc. or its subsidiaries in the United States and other jurisdictions. All other company and product names may be trademarks or registered trademarks of their respective companies. © 1996-2008 WebTrends Inc. All rights reserved.
ii WebTrends Marketing Lab 2
Sales and General
WebTrends, Inc. 851 SW 6th Ave. Suite 1600 Portland OR 97204 Phone: 1-503-294-7025 Fax: 1-503-294-7130 US Toll Free: 1-877-WEBTRENDS (1-877-932-8736) Email: firstname.lastname@example.org
Support and Services
Direct Technical Support: North America and Caribbean: 1-503-223-3023 Central and South America: 1-503-223-3023 Asia Pacific, Australia, New Zealand: 1-503-223-3023 Europe, Middle East, Africa: +44 (0) 1784-463-555
WebTrends Home Page: http://www.webtrends.com WebTrends Support Center: http://www.webtrends.com/support WebTrends Professional Services (Consulting and Training): http://www.webtrends.com/services Online Customer Center: http://www.webtrends.com/resources
About This Book and the Library
This guide provides an introductory conceptual overview of web analytics, supplemented with examples, graphics, and practical worksheets to help you understand WebTrends architecture and create a strategy for customizing WebTrends Analytics for your key business metrics. Topics covered in this guide include collecting web activity data, understanding visitor behavior, filtering and analyzing your data, measuring acquisition, conversion, and retention, and integrating web analytics data with other business data.
This book provides information for business users who are working with a WebTrends administrator to develop an approach to web site measurement.
Other Information in the Library
The library provides the following information resources:
Provides context-sensitive information and step-by-step guidance for common tasks, as well as definitions for each field on each window.
WebTrends Administration User's Guide
This guide provides complete information for using WebTrends Administration to set up and customize core Marketing Lab operations such as data collection, analysis, report content and style, and visitor session tracking. It includes conceptual and procedural information about features such as custom reports, data filtering, scenario analysis, and Express Analysis; assistance with common administrative concerns such as job scheduling and table limiting; and reference information such as the WebTrends Query Parameter Reference.
WebTrends Analytics Reports User’s Guide
This guide provides users who primarily use WebTrends Analytics Reports with the information they need to navigate, customize, save, and export reports and report data. .
Users who only have View Reports permissions automatically use WebTrends Analytics Reports instead of WebTrends Administration. While they can view the reports, they may not have access to any of the other controls. We recommend distributing the WebTrends Analytics Reports User’s Guide to these users as a introduction to navigating WebTrends reports and report data.
WebTrends Marketing Warehouse Software User's Guide
This guide includes information about using WebTrends Visitor Intelligence for ad hoc data analysis, using WebTrends Explore to analyze web business events by segment, and using WebTrends Score to identify qualified users based on their web site actions. It also provides a detailed reference to Visitor Intelligence report data. WebTrends administrators can also find information about installing, implementing and using WebTrends Marketing Warehouse.
WebTrends Marketing Warehouse On Demand User's Guide
This guide includes information about using WebTrends Visitor Intelligence for ad hoc data analysis, using WebTrends Explore to analyze web business events by segment, and using WebTrends Score to identify qualified users based on their web site actions. It also provides a detailed reference to Visitor Intelligence report data. WebTrends administrators can also find information about implementing and using WebTrends Marketing Warehouse.
ivWebTrends Marketing Lab 2
WebTrends Analytics Software Implementation and Maintenance Guide
WebTrends Analytics On Demand Implementation Guide
WebTrends SmartSource Data Collector User's Guide
WebTrends SmartView User’s Guide
A guide to installing and using SmartView and configuring WebTrends to work effectively with SmartView reporting.
WebTrends SmartReports User's Guide
A guide to using WebTrends SmartReports with WebTrends Analytics reporting for powerful data integration and analysis in the Microsoft Excel environment.
WebTrends Marketing Warehouse Schema Reference
Provides an overview of the Marketing Warehouse databases for experienced database administrators. This guide helps you understand the data in the Marketing Warehouse, giving you the foundation you need to use the data productively. It provides instructions for populating the Marketing Warehouse databases using WebTrends Administration and for viewing the data once it is available. It also describes how the databases are constructed and how that affects the function of the different types of data.
WebTrends Marketing Lab Programmer's Reference
This guide provides conceptual, procedural, and referential information that allows experienced programmers to customize WebTrends Marketing Lab data collection and reporting. It provides instructions for using the WebTrends ODBC Driver to query both the Marketing Warehouse and the WebTrends Analytics Report databases. It also includes documentation for the Active X, C, and Post Plug-Ins that can communicate with WebTrends Analytics.
WebTrends Visitor 360 Web Services Developer's Guide
This guide for WebTrends On Demand customers provides development information for using Visitor 360 web services to populate a Marketing Warehouse with off-site data, query and retrieve Analytics Reports data, and execute named queries against Marketing Warehouse data.
WebTrends Guide to Web Analytics
This guide provides an introductory conceptual overview of web analytics, supplemented with examples, graphics, and practical worksheets to help you understand WebTrends architecture and create a strategy for customizing WebTrends Analytics for your key business metrics. Topics covered in this guide include collecting web activity data, understanding visitor behavior, filtering and analyzing your data, measuring acquisition, conversion, and retention, and integrating web analytics data with other business data.
Your comments are very important to us. Please take the time to let us know about your WebTrends experience by doing one of the following: • Click Customer Center in the upper right corner of the WebTrends Marketing Lab banner. Then click Contact Us and click Submit Product Feedback in the right pane.
• From WebTrends Analytics Reports, click Help > Feedback from the upper right corner of the report. The Feedback page of the WebTrends web site opens in a new browser window. You can use it to report a bug, request a feature, or give general feedback about your user experience.
The Customer Center
The WebTrends Customer Center brings together a wide variety of materials to help you learn to use WebTrends Analytics more effectively, including white papers, interactive training modules, How Do I? Guides, and business case studies. To access the Customer Center, click Customer Center in the upper right corner of the WebTrends Marketing Lab banner.
The library uses consistent conventions to help you identify items throughout the documentation. The following table summarizes these conventions.
• Window and menu items • Technical terms, when introduced • Book and CD-ROM titles • Variable names and values • Emphasized words • • • • File and folder names Commands and code examples Text you must type Text (output) displayed in the command-line interface
viWebTrends Marketing Lab 2
Table of Contents
Introducing Web Analytics 1 The Purpose of This Book ......................................................................................................1 Who Should Read This Guide? ..............................................................................................1 The Current State of Online Marketing ...................................................................................2 What is Web Analytics? ..........................................................................................................2 Developing Intelligence About Web Customers ...............................................................3 How This Guide Fits with Your Strategy .................................................................................4 Measurable Improvement Cycle .............................................................................................5 Problems That You Will Solve ................................................................................................7
Defining Critical Metrics 11 Your Site’s Higher-Level Goals .............................................................................................11 Your Site’s Specific Objectives .............................................................................................11 Your Site’s Business Metrics ................................................................................................13 Content Sites .................................................................................................................14 Commerce Sites ............................................................................................................15 Lead Generation Sites ...................................................................................................16 Self-Service sites ...........................................................................................................17 Intranet Sites ..................................................................................................................17 Branding Sites ................................................................................................................18 Integrating Metrics ................................................................................................................18 Summary ..............................................................................................................................19 Objectives and Critical Metrics Worksheet ...........................................................................20
Collecting Web Activity Data 21 Understanding Data Collection Methods ..............................................................................21 Using Client-Side Tagging ....................................................................................................21 Using Web Server Logs ........................................................................................................23 Combining Web Server Logs and Client-Side Tagging ........................................................28 Hosted Versus Installed Software Solutions .........................................................................28 Summary ..............................................................................................................................28 Data Collection Worksheet ...................................................................................................29
Visitor Identification 31 Defining Web Activity ............................................................................................................31 Determining Unique Visitors .................................................................................................32
Sessionizing Your Visits .......................................................................................................32 Visitor Identifiers ...................................................................................................................33 Client IP Address or Domain Name ...............................................................................34 Combining IP Address and Agent Information ...............................................................35 Cookies ..........................................................................................................................35 Session IDs or IDs Embedded in URLs .........................................................................39 Authenticated User Name ..............................................................................................39 Summary ..............................................................................................................................39 Finding the Features in WebTrends Products ......................................................................40 Visitor Identification Worksheet ............................................................................................41
Defining Behaviors 43 Focusing the Scope of Analysis ............................................................................................44 URL Classification ..........................................................................................................44 WebTrends Methods of URL Classification ...................................................................45 Other Site Structure Issues ............................................................................................52 Summary ..............................................................................................................................55 Finding the Features in WebTrends Products ......................................................................55 Defining Behaviors Worksheet .............................................................................................56
Filtering and Analyzing Your Data 57 Profile-Level and Custom Report Filters ...............................................................................57 Include and Exclude Filters ...................................................................................................57 How Filters are Processed ....................................................................................................58 Setting Up Your Profile—Initial Filtering ...............................................................................58 Hit and Visit Filters ................................................................................................................58 Hits .................................................................................................................................58 Visits ..............................................................................................................................59 Hit Filter Criteria .............................................................................................................59 Visit Filter Criteria ...........................................................................................................65 Custom Reports ....................................................................................................................68 Parent-Child Profiles ......................................................................................................69 Summary ..............................................................................................................................70 Finding the Features in WebTrends Products ......................................................................70 Filtering Worksheet ...............................................................................................................71
Acquisition Metrics 73 What the Business Person Wants to See .............................................................................73 Performance Dashboards ..............................................................................................74 Entry/Landing Page .......................................................................................................74 Collecting the Right Data ......................................................................................................76 Referrers ........................................................................................................................76 Ad Campaigns ...............................................................................................................77
viii WebTrends Marketing Lab 2
Search Engines ..............................................................................................................81 Email Marketing .............................................................................................................83 Summary ..............................................................................................................................85 Finding the Features in WebTrends Products ......................................................................85 Acquisition Metrics Worksheet ..............................................................................................86
Conversion Metrics 87 Understanding Navigation Measurement .............................................................................88 Path Analysis .................................................................................................................88 Scenario Analysis ..........................................................................................................90 Internal Search .....................................................................................................................92 Exit Page and Exit Ratio Analysis .........................................................................................92 Visit-to-Exit Ratio ...........................................................................................................93 Dead-End Paths ...................................................................................................................93 Gleaning Demographic Information Through Registration Forms ........................................94 Evaluating Visitor Behavior by Browsing Your Site ..............................................................95 Summary ..............................................................................................................................96 Finding the Features in WebTrends Products ......................................................................97 Conversion Worksheet .........................................................................................................97
Retention Metrics 99 Visitor Segmentation and Behavior Segmentation ...............................................................99 Lifetime Value .....................................................................................................................101 Visitor History ......................................................................................................................101 Custom Segmentation ........................................................................................................104 Unique Visitors, Unique Buyers ..........................................................................................105 Finding the Features in WebTrends Products ....................................................................105 Retention Worksheet ..........................................................................................................106
Integrating and Exploring Data 107 Data Integration and a Web Data Warehouse ....................................................................107 Tying Data to External Databases ...............................................................................108 Reporting from a Web Data Warehouse ......................................................................110 Deeper Reporting and Exploration .....................................................................................110 Drilldown Capability .....................................................................................................111 Data Exploration ..........................................................................................................112 Overhead and Monetary Costs ....................................................................................114 Using Reports for Continuous Improvement ................................................................114 Data Integration and Exploration Worksheet ......................................................................115
Glossary ..................................................................................................... 117 Index ........................................................................................................... 138
WebTrends Marketing Lab 2
Introducing Web Analytics
Like any integral part of your business that requires dedicated time, money, and employees, your web site needs to prove its worth. You need more information such as who is visiting your web site, which web pages they are visiting, the order of web pages they are visiting, and which pages they are ignoring. Fortunately, the answers to these questions are available through a technology called web analytics. But web analytics is more than just a sophisticated software package and some hardware that runs it. You will need to apply a fair amount of thought and work to implement and make effective use of web analytics to improve your web site.
The Purpose of This Book
This book helps to demystify the mechanics of web analytics, removing the barriers that have long kept organizations from reaping the benefits of the solutions provided by web analytics. This book discusses: • • • • How to collect web traffic data How to set up your web analytics solution to give you the answers you need How to work with your software to get optimal performance with your web analytics What to consider when setting up your organization to run web analytics
These topics cover most of what any organization needs to know when choosing and implementing a web analytics solution. An in-depth discussion about these topics will give you an overall understanding of the field of web analytics and will help you initiate the process of analyzing your web site. By reading this book, you will obtain a comprehensive overview of all the options you have, which lets you make the choices that best suit your organization’s needs. You will also find a “Finding the Features in WebTrends Products” section in chapters 4 through 11 that will link many of the topics discussed in each chapter to WebTrends products. As an additional benefit, worksheets with pertinent questions are provided at the end of chapters 2 through 11 to help you in your quest to find the right web analytic solution. Also, please consult the “Glossary” on page 117 for a brief explanation of many terms used in this book.
Who Should Read This Guide?
You should read this guide if you have purchased or are considering purchasing WebTrends products and you manage or are tasked with making these products work with your company’s web servers, customer relationship management databases, and other decision-making support tools. This guide contains a wealth of technical information that you will not find anywhere else. If you work in marketing, product management, business development, sales, or related fields, you might not be interested in many of the technical details involved in setting up and using WebTrends products, but you probably do want to know how to use web analytics to achieve insight into who your customers are, what they do on your site, and what goods, services or information they want from you. This guide discusses what kind of information you can get from web analytics, and how you can fit that information into the larger context of your e-commerce efforts.
The Current State of Online Marketing
The online advertising arena has grown rapidly over the last few years to the point where current expenditures are estimated at $10 billion per year. Experts who watch business trends expect this amount to continue to increase exponentially. In fact, Piper Jaffrey recently projected online advertising spending to exceed $55 billion by 2010. The primary reason for this dramatic growth is the fact that although consumers spend more than 30% of their time online, only 5% of advertising dollars are spent online. Over the next few years, the gap between online advertising spending and online advertising consumption will close as more advertisers seek newer and better ways to reach their audiences. The following list describes some of the key individuals within marketing organizations and the primary interests that they have in online marketing: • • • • • Marketing executive asks if the company is reaching performance goals and where should the company invest its budget. Web producers ask how is the web site converting visitors, and how to make it easier for visitors to find what they want. Merchandisers and product managers ask how sales are performing and which changes to the product mix will increase revenue. Online marketing managers ask if the lead targets have been reached and how to improve the cost per lead. Customer marketing asks how email campaigns are performing and which segment will take action.
What is Web Analytics?
Web analytics is a tool that can help marketing professionals (in the previous section) to find the answers to their questions. But in general, web analytics means different things to different people. Consider the following examples: • To an executive, web analytics can help to determine if the web site has been worth the financial investment. Does the site produce results (defined at a high level) and are these results improving over time, especially after a redesign? To product managers, web analytics can help to reveal customer interest in an array of products and, consequently, affect product offering and pricing. To an IT manager, web analytics involves determining how much traffic the site experiences so that he or she can ensure that web servers can deliver web content flawlessly. To a technical support person, web analytics involves discovering whether a new series of online technical papers reduced customer support calls on a particular topic. To a marketing professional, web analytics means finding out whether ad space purchased on an external site was actually effective. To a web site programmer, web analytics means understanding which browsers and browser versions most visitors use so that the site can be designed to work optimally in those versions. To a web content developer, web analytics is discovering traffic patterns that influence his or her design improvements. To a sales person, web analytics is tracking which individual customers and prospects have been visiting the web site in order to narrow the sales approach for a given customer or prospect.
• • • • • • •
2 WebTrends Marketing Lab 2
Yet these perspectives are actually the applied definition of web analytics. The mechanics of web analytics are a little different. From a mechanics perspective, web analytics is a three step process in which you: 1. Collect web activity data. 2. Analyze the data that interests you. 3. Create meaningful reports on that data. The catch is that you can accomplish these three steps in many different ways. In the end though, each method arrives at a similar place—reports that help you determine whether your web site or a part of your web site is meeting its objective. But why is web analytics so frequently misunderstood? According to a Forrester Research report, only 23 percent of companies use web analytics to improve their online operations. The reason for this low turnout in the web analytics field is most likely because the basic concepts of web analytics and its implementation have never been fully discussed. web analytics is often viewed as black magic that only a few, gifted individuals know how to perform. In fact, many organizations have web analytics applications but experience so much frustration when using them that they abandon them altogether. Still other organizations find that the solutions they chose are either not comprehensive enough or are too comprehensive for their needs.
Developing Intelligence About Web Customers
By using WebTrends, you can develop more sophisticated and customer-centric information about your customers. The following graphic shows how this intelligence can lead you on a path from vague, general statistics to a sharp picture of who your customers really are.
Introducing Web Analytics
How This Guide Fits with Your Strategy
The overall strategy for your web site probably involves a combination of quantitative, data-driven, “factual” approaches along with subjective judgments, gut feelings, and emotional reactions. Equally, in your efforts to improve your web site, it is important to combine the “soft” approach with the “hard” to get the best results. This means involving a range of people in coming up with site enhancement propositions. From a strategy perspective this guide discusses how to measure and analyze your web site data, the results of this analysis then feeds back into the soft and hard sides of your strategy, which then drives recommendations and improvements to your web site. The following graphic shows an overview of how this guide relates to your overall web site strategy.
As part of your web site strategy, you need to identify the following: • • • • • The primary goals of your organization The primary goals of your site Goals of individual sections of the site Successful visit profiles The drivers to successful visits
4 WebTrends Marketing Lab 2
WebTrends Consulting and Training
Your strategy may require the help of WebTrends Consulting and Training so that your organization can implement, manage, and understand your WebTrends solution. WebTrends has expert consultants and trainers to help you meet the business requirements of your organization. WebTrends consulting engagements are focused on helping you implement your WebTrends solution and enabling you to manage WebTrends successfully on a day-to-day basis. Specialized training courses help your organization explore what WebTrends has to offer and gain a common foundation and understanding of how WebTrends applies to you. Together, WebTrends Consulting and Training helps you make the most of your WebTrends solution.
What You’ll Get:
• • • Consulting and training from experienced web analytics industry experts Faster return on investment and reduced time and resources required by your organization Valuable knowledge transfer on how to manage WebTrends successfully
The following graphic demonstrates the phased approach that WebTrends recommends to optimize your use of the WebTrends reports. Defining your eBusiness strategy and performance metrics is a key starting point. With key metrics and reporting requirements defined, a reporting solution is then implemented which provides information critical to guiding and strengthening your eBusiness strategies.
Measurable Improvement Cycle
You can implement a simple process that will help you to improve your web site by following a few proven steps. This process is called the Measurable Improvement Cycle, and creates a continuous improvement loop in which efforts are repeatedly refined through measurement.
Introducing Web Analytics
The following graphic shows the Measurable Improvement Cycle.
Applying this process to all web site decisions will help you focus your benchmarks and make critical adjustments to your web site, helping you to improve each time you complete the cycle.
Stage 1: Report
Report on the key metrics for each of your site’s objectives: • • • • Define the measurements you need. Configure your analysis solution and web site as per your measurements. Process and assemble site’s raw data into analysis reports. Provide analysis reports to appropriate department and individuals as needed.
Stage 2: Analyze
Use WebTrends to determine the performance of key metrics and site goals. Analysis in the form of reports allows you to: • • Set baseline performance. Evaluate the impact of site changes.
Stage 3: Decide
Determine what to do based on what the measurements tell you. Decisions might involve: • • • • Changing your web site. Altering marketing efforts. Revising content strategy. Updating your business model.
Stage 4: Act
Armed with the tables and graphs of your reports, you can optimize your site to improve performance of key metrics.
6 WebTrends Marketing Lab 2
Modify your web site according to your data. For example, you might adjust the steps in the shopping cart scenario. Remember that small incremental improvements are the goal. Try A/B testing. On the web this means that you are sending 50% of your traffic to one page and 50% of your traffic to another page. However, A/B testing may result in a reduction of the desired action that you want from your visitors—such as registering or purchasing. Filter x% of traffic to test against as an alternative to A/B testing. Just divert a small percentage of visitors to the alternate web page that you want to test. This may allow you to gather more accurate testing results. Perform usability testing on the changes you made to your web site.
Stage 5: Visitors React
The visitors to your web site may behave differently than you expected. For example, in modifying your shopping cart scenario, you may have caused some visitors to drop out of the process. You respond by measuring their reaction.
You will experience more success as you keep with the improvement cycle. Effective incremental changes involve a process rather than an end-result. Sometimes you may need only to change one or two things before you do another analysis. Incrementally refining your changes might help you more than making wholesale alterations.
Problems That You Will Solve
This section looks at some sample problems that you might want to solve and directs you to sections of this manual that contain relevant information.
Web Site Goal: Sell More Products Online
The following concepts allow you to understand who is looking at your products and buying them—or abandoning the buying process and when that happens. To sell more products online, you want to streamline the navigation through your site, so that people can see the products and offers that you intended for them. • Path analysis Path analysis will tell you if people are easily navigating to your products or if they are showing some confusion in getting to your products. For more information, see “Path Analysis” on page 88. • Scenario analysis A more specialized type of path analysis is scenario analysis. This type of analysis helps you discover if people are visiting all the pages in a scenario that you intended for them. For example, you can analyze a checkout sequence to see whether people complete the sequence or abandon it. One typical problem that scenario analysis helps to identify is when shipping information is only available within the checkout process. In such cases, you’ll see a high number of users abandon the checkout process after the page showing the shipping charge. Many of these customers were simply browsing to compare shipping charges with the competition.
Introducing Web Analytics 7
For more information, see “Scenario Analysis” on page 90. • Filters Filtering allows you to understand which segments of people are looking at your products and buying them. For more information, see “Filtering and Analyzing Your Data” on page 57.
Web Site Goal: Find Resellers for Products
If you are a company that manufactures products such as clothing, you can use web analytics to help you identify resellers for your products. • Registration location and scenario analysis If your web site has one or more pages with a special link or button that allows visitors to sign up as resellers, then you can do two things. 1) You can vary the location of the registration link or button on the web page to determine if you get more clicks on it, because of its location. 2) You can use scenario analysis if the registration process has a sequence of pages. If potential resellers abandon the registration process at a certain point, then perhaps it is too complicated, and you may need to simplify the process. For more information, see “Scenario Analysis” on page 90. • Filters If you are looking for a reseller in a certain part of the world, then you can filter your web traffic based on geography. For more information, see “Filtering and Analyzing Your Data” on page 57.
Web Site Goal: Distribute International Leads
If your company is getting online sales leads from many parts of the world, you will want to distribute those leads to the appropriate salespeople. • Content groups You could look at products that can be grouped together because they are of a similar type and then look at the people who selected those products. You might find that some products are being heavily selected from a certain part of the world and then assign those sales leads to the appropriate salespeople. For more information, see “Content Groups” on page 46. • Filters You can filter web traffic based on geography. For example, you might look at sales opportunities that came from the United Kingdom and simply forward those leads to your UK salespeople. For more information, see “Filtering and Analyzing Your Data” on page 57. • Custom reports You may want to develop custom reports after compiling information about visitor history and/or looking at registration information. The resulting custom reports can be tailored to the needs of your sales teams. For more information, see “Visitor History” on page 101.
8 WebTrends Marketing Lab 2
Web Site Goal: Sign Up for Newsletter
Companies that have a newsletter can use a variety of tools to track how effective they are in getting people to view and sign up for that newsletter. • Ad views and clicks. If you want to find out how many visitors have viewed or clicked on the link to your newsletter, you should understand the concepts of “ad views” and “ad clicks.” For more information, see “Advertising Views” on page 51. • Reverse path analysis You can see what route visitors took to get to your newsletter. You can determine which sections or pages are so inspiring that visitors decide they want to stay in touch with you. Also, using path analysis, you can see where those visitors go after viewing the newsletter. For more information, see “Path Analysis” on page 88. • Parameter Analysis If you allow visitors to sign up for a variety of new topics (like a graduated opt-in), you could use parameter analysis to report on topics in which visitors are most interested. Additionally, you could correlate those topics of interest with other web site activity, such as Content Groups. • Scenario analysis If you have a specific set of steps that you want your visitors to take, and one of those steps (such as in a checkout sequence) offers the visitor an opportunity to sign up for your newsletter, then you will want to use scenario analysis to determine if the offer is placed in the correct step of the sequence. If visitors abandon the sequence at the point in which they should sign up for your newsletter, then perhaps the web page needs to be designed differently. For more information, see “Scenario Analysis” on page 49. • Content groups You might want to find out what product or set of products have been visited the most over the past few months and then make that product or product set a centerpiece of an upcoming newsletter. To find out how groups of products are faring, you’ll use a concept called content groups. For more information, see “Content Groups” on page 46.
Web Site Goal: Optimize for Search Engines
Search engines are often a catalyst that drive visitors to your web site. They play an increasingly important role in the web environment. • Search engine analysis Examine search phrases and keywords to see what words are bringing visitors to your site to learn whether your site is getting traffic from all the terms that you expect them to use. The results will let you take action on your weak keywords. Which search engines are the most successful and least successful? You might also want to evaluate the quality of the traffic that the search engines brought to the site. Did various conversions occur? Did visitors spend a lot of time on the site? How many calls to action have been followed? For more information, see “Search Engines” on page 81. • Ad campaigns
Introducing Web Analytics
If you set up an ad campaign—which is tied to a specific search engine—as a referrer, landing page, or landing page parameter, you can examine how effective that campaign is. This could help you to determine which “paid” search engines are most effective. Which ad campaigns are the most successful and least successful? You might also want to evaluate the quality of the traffic that the ad campaign generated. Did various conversions occur? Did visitors spend a lot of time on the site? How many calls to action have been followed? For more information, see “Ad Campaigns” on page 77. • Spider and robot report You can determine how much of your raw traffic is attributed to spiders, which ones are indexing your site, and how deep in your site they are going. Spiders and robots are automated programs that crawl through the Internet to collect and index information, usually on behalf of a search engine or a monitoring company. You can use the report analysis to block spiders and robots from your web site.
Web Site Goal: Increase Customer Retention
Using web analytics, you can determine how well your web site is retaining customers. Consider these concepts: • New vs. returning visitors Learn about your new visitors and repeating visitors. Find the conversion rate of new vs. returning visitors. For more information, see “Determining Unique Visitors” on page 32. • Visitor behavior—frequency, recency, and latency Identify which visitors frequently return to your web site, how quickly they return to your site, and how much time elapses between visits. Once you’ve understood your visitor’s behavior you can present them appropriate advertising and thereby increase their monetary value. Understanding the visit cycle length can also influence how often you change the look of your home page, rotate featured products, and add new products. For more information, see “Visitor Segmentation and Behavior Segmentation” on page 99. • Path analysis Compare the navigation of visitors who purchase products to those who do not, and then fine tune your web site according to what you’ve learned. For more information, see “Path Analysis” on page 88.
10 WebTrends Marketing Lab 2
Defining Critical Metrics
Your Site’s Higher-Level Goals
Every web site has primary goals. Some sites sell products or provide information. Others specialize in games to provide entertainment. For example, the ultimate goal of a self-service web site is to save costs associated with other methods of customer service (such as email and technical support) versus generating revenue. Many sites serve a combination of purposes. For example, a symphony orchestra’s web site typically provides information about the organization and sells tickets to its concert season. Web sites for large companies often consist of individual sections with differing goals for each section. For example, one section might contain a series of pages devoted to commerce while another section offers customer service and a link to another section for lead generation. Keep in mind what constitutes a successful visit to your site. For a commerce site, this usually means that a visitor purchased a product online. And that visitor probably went through several steps—as defined in a shopping cart scenario—to complete the purchase. How do you improve your web site? What would make more visitors buy your products, play your games, complete the lead generation questionnaire? After you define what you want to improve, you can look more closely at your site’s objectives.
Your Site’s Specific Objectives
Although web sites are unique entities that serve a variety of purposes, some objectives are considered universal to nearly every site: • • • • • Increase visitor satisfaction - making site more convenient and valuable to visitors Decrease acquisition costs Increase conversion rates Improve customer/visitor retention Increase your web ROI.
However, since no two web sites are alike, each site can have individually tailored objectives. The following table identifies several types of web sites and some corresponding objectives.
• Increase sales and generate revenue • Complement offline channels • Increase average order size
• Research products • Buy products
Web Analytics Focus
• Buying & research behavior • Obstacles to purchase • Visitor-to-buyer conversion • Abandonment analysis • Campaign effectiveness • Purchase drivers • • • • • Research behavior Visitor-to-lead ratio Lead quality & cost Campaign effectiveness Call to action optimization
• Generate quality leads • Increase sales opportunities
• Research products/ services • Collect more information • Contact a representative • Find information • Conduct research
• Distribute information • Enhance marketing and service • Reduce costs
• Info-seeking behaviors • Ease of use and success • Electronic vs. traditional costs • Other supporting goals • Ad tracking, sponsorships, etc. • Frequency, depth, and length of visits • Popular audience interests for targeting/ segmenting • Ad tracking, sponsorships, etc. • Conversion from “entertainment” visits to other “revenue” or “branding” behavior visits • Frequency and quality of visits (are they an ad clicker?) • Advertising revenue generated • Visitor interest in content and preferences for segmentation • Audience growth, loyalty, engagement
• Develop audience loyalty • Monetize through ads or commerce • Brand building
Portals & Media Sites
• Generate revenue through ads, referrals, paid search placements, visitor services • Build loyalty • Increase page views per visit • Increase visit frequency • Subscriptions to magazine, newspaper, and online publications
• Quickly and easily find information • One-stop information source
12 WebTrends Marketing Lab 2
• • • • • Provide service online Reduce service costs Speed resolution rate Offer problem resolution Offer knowledge base information
Quickly and easily find answers to resolve issues
Web Analytics Focus
• Visit frequency and duration • Issue resolution rate • Tracking of email inquiries after reviewing help pages • Most successful type of help content/pages • Visit frequency and duration • Most popular content/ pages • Completion of a series of steps (scenario) • Visitor/department-level activity.
• Leverage Knowledge Base • Streamline operations • Provide access to critical applications
Quickly and easily perform duties
Of course, most sites have multiple objectives and consequently fall into several of the above categories. Businesses generally focus on more than just one task. For example, a company selling products will be concerned about customer service and lead generation for higher-end products. Also, large companies with multiple divisions may share portions of a web site and have numerous objectives. The message is clear: you must look at the chief characteristics of your web site. What does your web site do? What are the handful of metrics that will tell you that you are successful?
Your Site’s Business Metrics
What are the metrics that show you whether or not you are achieving the goals for your site? You need concrete measurements to know what you can improve. Regardless of your specific site objectives, you’ll want to measure the conversion rates of some scenarios (or steps through your site) to get a high-level view of your site’s effectiveness. For example, a commerce web site will probably examine the percentages of visitors that: 1. Visit your shopping section. 2. View a product. 3. Add a product to the shopping cart. 4. Start checking out. 5. Finish checking out. A customer self-service web site may be interested in the percentages of visitors that: 1. Log in to members page. 2. Visit various pages with pertinent topics. 3. Print or download information. 4. Log out.
Defining Critical Metrics
By measuring the visitors in each step of a scenario, you can determine where in the process you are losing the most people and then take action to improve the situation. The following subsections discuss metrics for several general web sites. The vast majority of web sites represent a combination of the following business models, as shown in the following graphic.
Content sites are media sites and specialty portals that are supported by sponsors and ads, subscriptions, premium services, and other means. Examples are Yahoo, CNN.com, Salon.com, and Consumer Reports. Content sites are typically interested in the following metrics:
Average page views per visit
Content sites desire an increasing amount of pages views per visit. By examining this metric in relation to content groups, you may gain more perspective on what areas are generating the most interest.
Average visits per visitor
How often are visitors returning each day, week, or month? This is an important metric that may indicate the success of a particular campaign.
Clickthroughs of on-site ads
Since many content sites are supported through advertising, monitoring the number of clickthroughs of these ads help you gauge the value of the ad.
First-time versus returning visitors
Does the content effectively engage visitors enough to make them return? By tracking the ratio between new and return visits over a period of time, you can determine if your site is attracting enough returning visitors.
Average visit frequency and recency
You will want frequency to be high and recency to be low to retain and grow your audience.
Content group activity and history metrics
If a content group experiences fewer and fewer visits, then you can investigate and take action.
14 WebTrends Marketing Lab 2
Number of search engine referrals
The number of visits referred by search engines is usually a critical metric for most content sites.
Specialized conversion rates
Conversion rates typically explore how many visitors move from one step to the next in a scenario that you are monitoring. Media sites may want visitors to register for topical newsletters to increase ad revenues and drive repeat traffic to the site.
Commerce sites are sites where companies sell their products and services. Examples are Amazon.com, WalMart, Converse, and Diamond.com. Commerce sites are typically interested in the following metrics:
Companies with high gross margins (gross revenue less cost of goods) have more money to spend on business operations such as research and development.
Gross margin return on Investment (GMROI)
GMROI is Gross Margin divided by demand creation expense for that order. That is, Gross Margin dollars are divided by the cost of the demand creation activity that drove the sale. This comes from being able to track the most recent campaign.
Represents the gross revenue minus taxes, interest, depreciation, cost of goods sold, and other expenses.
Represents the total invoice value of sales, before deducting for customer discounts, allowances, or returns.
Average order size
Represents gross sales divided by the number of orders—this reveals the average amount spent on each order. The higher the average amount, the better you are at motivating buyers to purchase more.
Accessory attachment rate
This the overall rate at which accessories are added to an order. This is the measurement of the number of orders which have an accessory attached to the order, divided by the total number of orders. This measurement determines how to grow the overall average order size, as well as growing the gross margin/profit of a single order. Accessories typically have the highest gross margin on a site and significantly increase the profitability of an order. For example, the cables on a DVD Player order may have as much profit dollars as the player.
Sales conversion ratio
Represents the ratio of visitors to sales and visits to sales.
Customer retention rate
Represents the number of repeat customers divided by the number of total customers over a period of time. Commerce sites strive for repeat business.
Cost per sale
Represents marketing expenses divided by the number of sales during a period of time. Low cost per sale means efficient marketing and a higher net profit.
Defining Critical Metrics
Customer acquisition cost
This is marketing expenses divided by the total number of orders from unique, first-time buyers over a period of time. If it costs a lot to acquire new customers, then you may have to retool your marketing effort.
Average lifetime value
What is the value of your customers over a period of time. Is it increasing?
Specialized conversion rates
Conversion rates typically explore how many visitors move from one step to the next in a scenario that you are monitoring. An example of a specialized conversion rate for a commerce site: your site invites visitors to register for a newsletter or sign up for a contest. Compare how many visitors see the offer with how many actually sign up.
Lead Generation Sites
Lead generation sites offer information for sales processes by actively “capturing” visitors as leads. This usually occurs after visitors register or contact a sales representative. Examples include B-to-C web sites such as autos and homes, and Business-to-Business (B2B) web sites such as Siebel, Peoplesoft, and Boeing. Lead generation sites are typically interested in the following metrics:
Visitor-to-lead conversion ratio
This represents the percent of visitors that register or otherwise become a lead over a period of time. If this metric dips or peaks, you should evaluate conversion rates by acquisition source (campaigns).
Total number of leads
If the number of leads does not grow, then a site may need to be re-evaluated. Consider examining the number of leads from search engines, campaigns, partners, or the number of leads for different products or from a geographic region.
Cost per lead
Represents marketing expenses divided by the number of leads generated during a period of time. This metric contributes to understanding the cost of marketing campaigns and collateral.
Lead close ratio
This is the percentage of collected leads that ended up closing as a sale. If leads are “closed” through channels other than your web site, you may have to track lead closure manually.
Average visits or page views per visitor
If your site is seen as a resource, it may attract more leads that value the content.
Marketing campaign conversion rate
This is the general effectiveness of campaigns at driving visitors to register as leads.
Specialized conversion rates
Conversion rates typically explore how many visitors move from one step to the next in a scenario that you are monitoring. An example of a specialized conversion rate for a lead-generation site: your site wants to evaluate which methods (such as a newsletter or a webcast) lead to the highest closure rates.
16 WebTrends Marketing Lab 2
Self-service sites focus on helping customers resolve issues and/or learn about uses of the product or service without the aid of human interaction. Self-service sites are often a component of another model but can stand alone. Examples are support/knowledge base sites of most manufacturers and software developers, and online banking. Self-service sites are typically interested in the following metrics:
Average visits per visitor
An increase or decrease of average visits per visitor may be seen as positive or negative, depending on the site’s objectives. On the one hand, an increase is good for a governmental web site or an intranet maintained for employees, because it shows that visitors are performing many tasks, such as scheduling vacations, reading corporate policies, or checking on 401K plans. On the other hand, a software manufacturer may want the visits per visitor to decrease, indicating that people are finding what they need quickly.
Average page views per visit
The same considerations apply here as with visits per visitor. Compare average page views per visit with content groups to know whether a decrease or increase in activity is good or bad.
Knowledgebase searches per visit
How easy is it for visitors to find the information they want? If some knowledgebase articles are searched quite often, you may have to put better explanations into your product.
Number of zero result queries
This represents how often a visitor searches on a term and receives zero search results. You need to add new content if visitors received zero results after querying the same or similar keywords.
Online resolution rate
This rate is the percentage of site visits that resolve issues online versus those that need additional help over the phone or email.
Percentage of total support requests handled online
This information helps to identify which support options visitors are using and to what degree. If a certain option gets more attention than others, then you might consider upgrading the corresponding part of your product.
Specialized conversion rates
Conversion rates typically explore how many visitors move from one step to the next in a scenario that is being monitored. An example of a specialized conversion rate for a self-service site: a cellular company might want to allow its customers to edit their general account information, modify their calling plans, or download new ring tones.
Intranet sites are primarily company or organization sites that provide service for employees. Employees typically use intranet sites to schedule vacation, to download and print medical forms, to check up on company policies, and a variety of other tasks. Intranet sites have a lot of the same issues as self-service sites except that you know your total number of visitors (the employees). Therefore, the resulting reports will accurately reflect usage in relationship to a known number of visitors.
Defining Critical Metrics
Intranet sites would use the same metrics as the self-service sites. For example, by using scenario analysis you could look at the steps in a process such as filling out a vacation request form. Perhaps you would find that some employees abandon the process at a certain step because they are still unsure about their vacation plans. This would be similar to the steps explored in the Specialized Conversion Rate mentioned in the metrics for self-service.
Branding sites are those that seek to promote interaction with visitors and engage them with a brand. Sponsored by companies, initiatives, and/or events, branding sites intend to generate buzz, interest in a product/company, or stimulate sales. Note that these sites do not justify their existence on sales/leads generated or ad revenue. Examples of branding sites are absolut.com, movie sites, and Coca-Cola. Branding sites are typically interested in the following metrics:
Monitoring unique visitors by day, week, month, quarter, and year helps to evaluate the effectiveness of your online branding.
Depth of exploration
This includes measures such as average page view per visit, length of time, and content group exposure. When tied to a campaign, you can find out to what “depth” that campaign affected visitors.
Successful branding sites attract multiple, continuous interactions with visitors.
Average visit frequency, recency, and latency by content area visited
These measurements continue the concept of sustained interaction with visitors. Loyal visitors, for example, are the ones that typically purchase more products.
Specialized conversion rates
The rate at which visitors play games, download coupons or screen savers, enter contests, and so forth, and then register with your site is very important.
In terms of online advertising, organizations are spreading their marketing dollars across multiple channel and partners, such as Google, Yahoo, MSN, Wall Street Journal Online, email providers, and many others in order to reach customers. The WebTrends Marketing Lab consolidates all of your performance data and customer profile information that is driven by these channels and partners into a single consistent metrics framework. The WebTrends Marketing Lab then processes this consolidated performance data and provides a wealth of valuable reports for you to take action on.
18 WebTrends Marketing Lab 2
The following illustration describes the integrated reporting solution that the WebTrends Marketing Lab provides.
WebTrends Marketing Lab goes beyond traditional web analytics by providing a more comprehensive Marketing Performance Management (MPM) solution. This integrated reporting system delivers top-line performance metrics, operational analysis, and advanced business intelligence functions. The following are just a few examples of what the WebTrends Marketing Lab can do for your organization. • You may have customers sitting on the “purchase fence,” but you are not sure who they are. WebTrends Marketing Lab allows you to easily find out which customers are most likely to buy and then populate a list of these customers that you can send to your email provider. You may want to increase customer lifetime value by means of an up-sell offer. You can use the WebTrends Marketing Lab to query your data and identify the customers who are most likely to buy your new product line. Then you can use this information to focus your selling effort to these customers. To improve your customer service, you can use the WebTrends Marketing Lab to identify valuable customers, who are struggling to find answers online. From the results of your query, you can create a list of customers who have problems and then execute a customer service campaign that will help them.
After your company has firmly determined the objectives for its web site and determined which specific metrics to track, you can use WebTrends to get the reports that you need. These reports will influence the way you change your web site. You might, for example, improve the content in a sequence of steps that leads to the purchase of an item. In most cases, it is best to make small, incremental changes to your web site. You can then direct WebTrends to measure your visitors and get a new set of results to study.
Defining Critical Metrics
Of course, after you’ve made your changes, you may need to re-examine your site’s goals and objectives, and then add a new set of measurements. This is part of the continuous Measurable Improvement Cycle that was discussed in Chapter 1 on page 5. To help you think through the objectives and critical metrics of your web site, you can refer to the “Objectives and Critical Metrics Worksheet” on page 20. To begin understanding how to collect the data that you will explore with web analytics, continue to the next chapter, Chapter 3, “Collecting Web Activity Data” on page 21.
Objectives and Critical Metrics Worksheet
Use the information you’ve just learned about high-level goals, specific objectives, and web site metrics to fill out this worksheet. Consideration
What are the high-level goals of your web site? What would a successful visit to your web site be? What business model is your site? (Commerce, Content, Self-Service, Lead Generation, Intranet, or Branding/Campaign) How would you improve your web site?
What are more specific objectives for your web site? • Business goals • Visitor goals What do you need to measure to improve your site?
20 WebTrends Marketing Lab 2
Collecting Web Activity Data
Understanding Data Collection Methods
Using Client-Side Tagging
The following graphic shows a typical client-side tagging process.
2007-03-04 00:08:18 proxy7.hotmail.com W3SVC3 web1 192.168.1.1 GET /ads/default.asp
redir=products&ad=http%3A// www.boatdealer.com&WT.mc_n=Boat%20Dealer%20Campaign&WT.mc_t=Banner&WT.mc_s=3/3/ 2001&WT.mc_c=60&WT.ad=P-32,%20P-58,%20P72%20Options%20Offer&WT.sv=Web%20Server%201&WT.ti=Advertising%20Redirect&WT.tz=420&WT. ul=en&WT.cd=32&WT.sr=1024x768&WT.jo=Yes&WT.js=Yes&WT.co=Yes 200 0 1 75 1 80 HTTP/1.1
Microsoft+Internet+Explorer/4.40.305beta+(Windows+95) WEBTRENDS_ID=192.168.16.1481615253808.29527727 http://www.boatdealer.com/dealers/pacific/dealerlist.htm
The italicized text contains client-side tagging parameters that were used to fetch the data from a database that populated the web page, default.asp.
WebTrends has developed special data collection parameters called SmartSource parameters. In the example above, all SmartSource parameters begin with “WT.”
Benefits of Client-Side Tagging
In general, the client-side tagging is extremely effective for attaining business metrics but not for examining the underlying web server behavior.
22 WebTrends Marketing Lab 2
• • • • •
Client-side tags capture data for only the pages you want to track. This method reduces the amount of data you have to store or process. Client-side tags act as an automatic filter, because they do not collect images and other kinds of hit data that you do not want to collect. This automatic filtering helps reduce the size of data storage. Client-side tags can be implemented on your web pages quickly. Client-side tagging avoids problems of co-located servers and content served from multiple sites. Because the script runs each time the page loads, you have accurate visit and page counts, even when pages are loaded from a caching or proxy server.
Drawbacks of Client-Side Tagging
• • • • • • • • With a software solution, client-side tags require additional hardware to run the data collection server. Client-side tags require time or technology to embed the script in each page you want to track. You can only track error pages by embedding the script in the tags. If a visitor’s browser is not enabled to run the scripts, you only collect page and visitor activity, not details about what was visited. If a redirect page does not contain the data collection tag, it cannot get tracked. This could be crucial if you are using redirect pages to track advertisements. If a page load is interrupted before the data collection script is run, the visit to the page does not get recorded. Because most spiders do not run the data collection script, visits from spiders are not captured. Without custom configuration, client-side tags only capture HTML pages (such as .htm, .asp, .html). Consequently, downloads are more difficult to track.
Using Web Server Logs
Each time a visitor views something on your web site, downloads a file from your site, or in some other way requests something from your site, the web server—which holds and delivers the content for your site— adds a record to a log file. This record contains some basic information about the request the visitor made. Some of this information is known directly by the server, such as the time, date, content requested, and the size of the content requested. Other information is obtained through a cooperative and heavily standardized relationship between the browser and the server, in which the visitor’s browser is programmed to send certain information, such as the IP address of the computer it is running on and specifics about the browser version and operating system of the visitor’s computer. Most web server log files are text files that contain the following pieces of information: • • Date and time that the visitor made the web server request Required for time-sequencing records and identifying paths. The IP address (Internet Protocol address) or domain name of the visitor’s computer Not required, but strongly recommended. This may be used for visitor tracking—to get the domain of the visitor—and for looking up geographical information. • The web server’s name—on your web site Not required; not used
Collecting Web Activity Data
The web server’s IP address—on your web site as seen from the outside world Not required; not used The method used in the request—such as GET, POST, and HEAD Not required, but it used for determining the type of action the visitor took, such as a page request or an upload.
• • • • • • • •
The URL of the requested content Required. All content-related information is derived from this field. Any query parameters, if additional information is needed Not required but strongly recommended. Used for analyzing dynamic content. The return code—successful or failed delivery of the request Not required. Used for reporting on user and system errors. The number of bytes sent by the web server to the client Not required. Used for reporting on bandwidth usage. The number of bytes sent by the client to the web server Not required. Used to report on the amount of data sent from visitors to the web site. The amount of time (in milliseconds) to fulfill the request Not required, but if present, this is used for reports involving server response time. The port on the client machine used to send requests and receive the requested data Not required. Not generally used. The client machine’s browser type and version number (also known as “the user agent”) Not required. This is used for determining which browsers are in use, and for recognizing various types of spiders and search engine robots.
Cookie information, if the client machine has a cookie for your site Not required, though very useful for tracking unique visitors. Also, cookies can contain other, sitespecific information, which can be analyzed and reported on.
Referrer information, if the visitor was sent to your site from an external site Not required. Used for recognizing how visitors arrived at your site, especially via search engines.
This logged information and the order in which it appears has been specified by the software contained in the web server that keeps the log files. For Microsoft systems, the software is called Internet Information Services (IIS). You can program the software to reorder or drop pieces of information that you might find unnecessary, but it is best to do this only after you have gained some expertise with web analytics. Each log entry appears as information on one very long line in the file. The following sample log entry has been split over several lines so that you can read it more easily:
24 WebTrends Marketing Lab 2
2007-09-16 00:01:58 188.8.131.52 W3SVC82 HERC 184.108.40.206 GET /products/thingamajigger.html 200 4199 363 266 80 HTTP/1.0 Mozilla/4.72+[en]C-SBI-NC472++(Windows+NT+5.0;+U) WEBTRENDS_ID=192.168.32.180-3425858080.29527895 http://www.awebsite.com/thingamajiggerad.html
The following graphic explains this log entry by relating each bulleted item above to the corresponding information in the sample log entry.
Your log file can vary from this example, because you can configure your server to include the information you want. Also, the information available may vary according to the brand of server software (for example, IIS, Sun Java System, or Apache). Please refer to the server software’s documentation for directions on how to activate logging. In cases where your content management system produces non-standard URLs, you may need programmers who can write scripts to preprocess log files before analyzing them with WebTrends.
Log File Rotation/Rollover
Because web server logs can grow quickly and consume disk space, you likely need to transfer them from that server to another storage unit on an ongoing basis. Transferring logs is also known as “rotation” or “rollover.” Whether you keep log files on the same server or transfer them for storage, you probably want to have data available for trending analysis. For many organizations, a transfer of web server logs might occur on a set interval—perhaps once a day— but if a site experiences enormous amounts of traffic, these log files can be rotated off the server even more frequently, perhaps every hour. After they are rotated off the server, a new log file begins. Each log file receives a name that makes it relatively easy to track. For example, the log file for a web site for October 3, 2007 might be called ex20071003.log, in which the naming format is year/month/day. Log files for a high traffic site can have daily files that can reach gigabytes in size. To save disk space, once a file is rotated off the server, it is often compressed with an application like PKZip or WinZip. Fortunately, because the log files may contain many repeated elements, such as dates, URLs, browser and browser versions, you can compress the log files down to 5 to 10 percent of their original size. For more information about log file rotation, see WebTrends Implementation and Maintenance Guide.
Collecting Web Activity Data
The following graphic shows an example of how log files are rotated off of your servers and placed in a zipped archive or database.
Log File Access
WebTrends Analytics software will need access to the location where web log data is stored. Most log files are either stored on a mapped network drive or on a remote server that may be accessed using File Transfer Protocol (FTP). You may need to provide WebTrends with the required user name and password authentication information. If you choose to import the log files on a regularly scheduled basis, be aware that log files imported using FTP or HTTP are brought over in their entirety. You cannot transfer only the first 10,000 lines of a log file or the last 3,000 entries. The following graphic shows an example of how log files are rotated from co-located servers.
26 WebTrends Marketing Lab 2
How frequently you import the log files depends on how much activity your site receives. As a general rule, most sites bring over their log files once a day. However, if your site has high levels of activity and generates extremely large log files, you may need to transfer files more frequently. This reduces the data volume that must be handled at any given time. WebTrends is designed to recognize which files have already been imported, and only brings in files that contain new data. In comparison, accessing your log files from a network drive is a common way of obtaining your log file data because WebTrends treats it as though the log files were stored locally. However, because the data comes across the network from the mapped drive, the analysis process is slightly slower.
One week’s worth of log file data gives you a snapshot of the volumes of activity on a site, but three months’ worth of data gives you real insight into trends. Once you understand the trends, spikes and anomalies become evident and usually their cause can be traced and evaluated.
Benefits of Web Server Log Files
In general, the benefit of web server log files is that they tell you about the mechanism of delivering web pages, and, with a bit more work, they provide business metrics. • • Most web servers generate log files, so they are typically easily and immediately available. You do not need to decide in advance exactly what data you want to report on. Web server logs allow you to go back to the raw data at any point and change what you want to analyze, as long as data was initially logged. Even when a server goes down, it does not lose the web server log data, because the data collection device and the server are one and the same. Log files capture all downloads and non-HTML files in addition to HTML files. You can report on IT-based metrics such as reports on spiders, downloads, bandwidth, load balancing, and errors.
• • •
Drawbacks of Web Server Log Files
• • • • • If an ISP hosts your site, you may not have access to your log files. Log files collect all activity, even data that does not interest you. This method requires more storage space than web activity files generated by client-side tagging. If a log file exists but WebTrends cannot read it, the log file might be corrupt. If log files are missing, you might need to look for them elsewhere on the system. If the web site is hosted on geographically disperse servers, WebTrends gathers all the log files to one place and requires a means of ordering the records from all the log files. It must then determine which hits are part of the same visit. If time stamps on the various web server logs are not in sync, results can be inaccurate. You must also have a way to handle server disruption, or the results can be inaccurate. Log files cannot record repeat requests when a page is accessed from a caching server. Inaccurate information because of proxy servers and content delivery networks, such as AOL, AT&T, and Earthlink. For more information, see “Proxy Server Buffers” on page 35. Depending on the level of sophistication, the web analytics software installation and configuration may take time. The learning curve for this software is sharp and steep.
• • •
Collecting Web Activity Data
You must maintain the equipment and software yourself unless an ISP does this for you. You must write scripts (or purchase software containing ready-made scripts) to handle non-standard URLs that may need more processing to understand correctly.
Combining Web Server Logs and Client-Side Tagging
Companies that analyze data from web server logs and client-side tagging have the best of both worlds. They can use the log files to get information about the web server activity—primarily IT-based metrics such as reports on spiders, downloads, bandwidth (for example, bytes delivered), load-balancing, and errors. They then use client-side tagging to get higher-level business metrics.
Hosted Versus Installed Software Solutions
After choosing whether to use web server logs or client-side tagging, you need to determine if you want to hire a service to do that for you (called a hosted solution) such as WebTrends Analytics On Demand, or if you would rather be responsible for collecting and analyzing all the data yourself (called a non-hosted solution) and purchase stand-alone software such as WebTrends Analytics Software. Using a hosted solution is an attractive option for several reasons. The foremost reason is that you don’t have to maintain the web analytics software or hardware, and that you might be able to write off the service as an operating expense. Also, a hosting service arrangement doesn’t require the additional setup time that complex software solutions require. If you don’t like the service, you can cancel or finish out the contract and disable the data collection. In contrast, installed software (non-hosted) solutions provide greater control regarding the data you can analyze and in the way you can present that data. With data collected from web server data files, the most common kind of non-hosted solution, you can store web activity data indefinitely in raw log file format or processed data in a web data warehouse. This means that at any time you can re-analyze the data, combine it with external data sources, or run deeper analyses using third party software. Another key advantage of installed software is privacy, because you control the data, which is never stored on a third party server. Privacy is especially important for financial industries such as banking and insurance businesses. The main drawback of installed software is that you must maintain the software and hardware associated with your analysis solution. For this reason, the expenses are viewed by accounting as company assets, which are only depreciable and not deductible. Traditionally, the client-side tagging model has been primarily used as a hosted solution with products such as WebTrends Analytics On Demand and web server data file analysis has most often been used with software (non-hosted) solutions. However, with the advent of data collection servers, organizations can now use client-side tagging to collect activity data themselves (as a non-hosted solution) and either report on that data directly, or store the data in a data warehouse.
Which data collection method you should use really depends on the method that best meets your analysis needs and budget. If you know exactly what data you want to analyze and you only want some basic web activity reports, using hosted client-side tagging may be the sensible choice. This method reduces the amount of data that you have to collect and minimizes web data activity file storage issues. For small businesses, the hosted client-side tagging is also the least expensive method that delivers basic reports such as Pages, Visitors, and Referring Site.
28 WebTrends Marketing Lab 2
On the other hand, if you think that you may want to shift your analysis approach down the road, and want to keep all your options open, collecting the web server activity data or using a non-hosted data collection server gives you far more flexibility. Some organizations choose to combine both web server log and client-side tagging methods. They generate standard reports using client-side tags or server side tags, but collect and store web server log data to allow flexibility later on. In the future, many organizations may find that using a client-side tagging solution more attractive than using web server data. Tagging solutions collect the same information that web server logs can, allowing in-depth and flexible analysis and reporting, yet offer immediate report generation on standard data.
Data Collection Worksheet
Use the following worksheet to understand how you want to collect data about your web site. Consideration
Need access to log files? (Note: Hosted services typically do not provide access to log files.) Need to keep data for an extended period of time to do comparisons? Capture information on all downloads (HTML and non-HTML files)? Use multiple or co-located web servers? All servers are available at all times? Can afford up-front investment in terms of capital and training time? Can maintain additional hardware equipment and software? Need to write off costs as an operating expense? Capture data of only specific web pages? Quick install/uninstall? Can afford extra hardware? Can embed code in each page to be tracked (also redirect pages)? Only care about HTML pages and business metrics (no interest in IT-based metrics)? Prepared for software costs related to licensing?
Collecting Web Activity Data
Have the people (IT) resources? Know what kinds of information needed (business and/or IT)? Have the storage retention/space (time/how long)?
30 WebTrends Marketing Lab 2
The main objective of web analytics is to understand how web visitors are using your site (what pages are visited and what actions are taken) so that you can determine if they are doing what you want them to do. • • • Are visitors responding to ads? Are visitors making purchases or downloading white papers? Are visitors reviewing your technical support materials rather than calling your technical support personnel?
These are questions that you can answer by using WebTrends. Your web activity data file, whether generated by the web server itself or collected and created by a WebTrends data collection server, can tell you more about the activity on your site. But how can you tie activity to individual visitors? How can you tell whether a hit to a product information page and a hit to the pages of a shopping cart were all done by the same visitor? If you knew that, you could say that a particular visitor read the product’s description, decided to purchase it online, and then completed all the steps required for making a purchase. Tracking visitor activity can be quite complex, so it is important to keep in mind that you will spend more time, effort, and resources as you strive for more clarity and accuracy in understanding who your visitors are.
Defining Web Activity
From a high level, web activity includes which areas of the web site were visited, which products were viewed, and which actions were taken with those products. Visitors typically go through a path of pages. After you determine what the actions were and who did them, you can derive meaning from the activity (presented in easy-to-understand reports from WebTrends) and take actions such as revising your web site or tailoring messages for special customers. The following terms are commonly used when discussing web activity.
Represents the person or agent that generates the visits. Agent indicates a program, such as a robot or spider that is used to visit web sites.
Denotes a sequence of a visitor’s page views up until the point at which the time between two successive page views or hits is greater than the timeout session length (usually thirty minutes). Visits are also called visitor sessions. Much marketing research focuses on statistics for visitor sessions as an accurate picture of visitor activity. For more information, see “Sessionizing Your Visits” on page 32. Note that if you modify the session timeout length in your web analytics solution, you get different session visit count. For example, shortening the timeout length increases the count in the number of visits.
Represents a hit to any file classified as a page (such as html, htm, php, and asp pages). For sites using frames, an actual page viewed may consist of several HTML documents.
Represents any individual item that is delivered from the server to the client. A single visitor action could result in dozens of hits. For example, when a web page is delivered to a client’s screen, it may arrive with graphics, icons, ads, sidebars with links, frames, and other items that all count as hits. Although the volume of hits is an indicator of web server traffic, it in not an accurate reflection of how much real information your visitors are looking at.
Determining Unique Visitors
In order to associate web site activity to the actual visitors who performed that activity, you first need to uniquely identify the visitor responsible for each activity. Once you have identified the unique visitor for each page view, you can group all of the page views from a specific visitor into a visit session. WebTrends does this for you by associating all of page views analyzed to the visitors responsible for them. WebTrends also lets you track returning visitors. This means that when a visitor comes back for a new visit to your site, you can associate that visitor with the visitor’s previous activity. By tracking what your visitors do over time, you can establish major trends in visitor behavior on your site. The key here is to distinguish one web visitor’s actions from all other visitors’ actions. By understanding what group of actions each unique visitor did, you can discern how visitors in general are using your site. Questions you can answer by identifying unique visitors include: • • • • • How many new visitors came to my site during a specific time interval? How many visitors came to my site from an ad campaign on a search engine like Google? How many returning visitors came to my site during a specific time interval? Is the majority of the activity coming from new or returning visitors? How much time are visitors spending on my site?
Sessionizing Your Visits
Sessionizing is the process of assigning a unique visitor to one or more actions that occurred within a defined time period, or visit. A visit denotes a sequence of page views or hits up until the point in which the gap between two successive hits is greater than the defined timeout session length (usually thirty minutes). The following example shows records from a typical web server activity file.
2007-01-01 00:12:12 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 20072007-01-01 00:19:59 18.104.22.168- W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:24:43 22.214.171.124 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:29:59 126.96.36.199 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:40:46 188.8.131.52 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:41:22 184.108.40.206 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:44:00 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:44:17 18.104.22.168 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 00:46:13 22.214.171.124 - W3SVC3 HERC 192.168.1.1 POST 2007-01-01 00:48:24 126.96.36.199 - W3SVC3 HERC 192.168.1.1 POST 2007-01-01 00:59:59 188.8.131.52 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 01:01:13 184.108.40.206 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 01:03:02 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 01:04:40 18.104.22.168 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 01:06:32 22.214.171.124 - W3SVC3 HERC 192.168.1.1 GET 2007-01-01 01:09:01 126.96.36.199 - W3SVC3 HERC 192.168.1.1 GET 32 WebTrends Marketing Lab 2
2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01 2007-01-01
01:09:18 01:10:51 01:11:30 01:14:48 01:17:06 00:29:59 01:19:52 03:19:59 03:21:02 03:23:29 03:25:34 03:33:55 03:39:59 03:43:08 03:59:59 04:00:00
188.8.131.52 - W3SVC3 HERC 192.168.1.1 GET 184.108.40.206 - W3SVC3 HERC 192.168.1.1 GET 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 18.104.22.168 - W3SVC3 HERC 192.168.1.1 GET 22.214.171.124 - W3SVC3 HERC 192.168.1.1 GET 126.96.36.199 - W3SVC3 HERC 192.168.1.1 GET 188.8.131.52 - W3SVC3 HERC 192.168.1.1 GET 184.108.40.206 - W3SVC3 HERC 192.168.1.1 GET 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 18.104.22.168 - W3SVC3 HERC 192.168.1.1 GET 22.214.171.124 - W3SVC3 HERC 192.168.1.1 GET 126.96.36.199 - W3SVC3 HERC 192.168.1.1 GET 188.8.131.52 - W3SVC3 HERC 192.168.1.1 GET 184.108.40.206 - W3SVC3 HERC 192.168.1.1 GET 220.127.116.11 - W3SVC3 HERC 192.168.1.1 GET 18.104.22.168 - W3SVC3 HERC 192.168.1.1 GET
If you look at the activity of 22.214.171.124 (remember that this is a visitor’s IP address), you will notice that it has two visits, which are separated by a gap of at least thirty minutes. In general, in order to sessionizing requires two basic elements: • • A time stamp, to determine the start and end of a visit and to order page views in a time sequence. A visitor identifier that ties each page view to the visitor responsible for the page view.
The time stamp requirement is easily handled because WebTrends SmartSource data collection servers and web servers include a time stamp for any page view recorded. As long as Greenwich Mean Time (GMT) is used to indicate the time, servers that are located in different time zones will not have any problem understanding the time sequence of the data. The more complicated requirement is the visitor identifier, which is discussed in the next section.
You have several different methods available to identify the visitor associated with web site activity. Some methods identify unique visitors strongly, while others can only weakly identify visitors. The following methods provide weak visitor identification: • • • • • • Client IP address or domain name Combination of IP address and agent information Cookie (persistent or session-only) Session IDs Data embedded in the URL Authenticated user
The following methods provide strong visitor identification:
These methods are listed in order of increasing accuracy. The order also corresponds with the complexity of your site management. At the very minimum, your web analytics solution can examine the client’s IP addresses. The next best method is the combination of IP and agent, but the very best method is a combination of authenticated users and another strong identification method. In other words, the IP address is easy to identify while the authentication of users is much more difficult. As a best practice, you should use first-party cookies to identify unique visitors because they provide a reliable method with little overhead.
Visitor Identification 33
Though each method has its strengths and weaknesses, you may encounter such issues as:
The ambiguity of the visitor identifier
If two visitors can have the same identifier at the same time, they will appear as a single visit by the same visitor.
The problem with aliasing of a visitor identifier within a single session
If a single visitor has more than one identifier (for example, an alias) within a session, that visitor will appear to be multiple visitors, each having its own visit session.
The problem with the persistence of the identifier across multiple sessions
If a single actual visitor has two different identifiers from one session to the next, that visitor will appear to be two separate visitors. This causes an inaccurate count of unique visitors and new versus returning visitors. It also doesn’t allow you to accurately accumulate a single visitor’s activity over the lifetime of that visitor. As we discuss the various methods for identifying visitors, you will recognize how each method has one or more of these three issues to contend with.
Client IP Address or Domain Name
The easiest method by which to identify unique visitors involves using the visitor’s IP address or domain name. The domain name is the text name corresponding to the numeric IP address.
Domain Name System (DNS) is the method that the Internet uses to convert difficult-to-remember numbers, such as 10.17.243.32, to easy-to-remember names, such as www.yahoo.com (which are easier to read and comprehend than a series of numbers). The reason for this conversion is because the underlying protocol for the internet, TCP/IP, uses difficult-to-remember numbers to connect to other computers. When a visitor comes to your site, either that machine’s IP address or the domain name of the IP address automatically gets recorded in the web server data activity file. Which of these two identifiers gets recorded in your web data activity file depends on how your web server is configured to log hits. They can be configured to perform Domain Name Service (DNS) lookups while logging entries, or they can be configured to simply record the IP address. Many web servers do not perform lookups while logging information because it slows down delivery of the web visitor's requested content. However, if IP addresses are not resolved during creation of the web data activity file, you can always perform a DNS lookup after the web data activity file has been created. One of the major benefits of using IP addresses and domain names to identify the visitor is that many DNS servers contain additional information about the IP address or domain name, such as the location and company. This tells you where your visitors are coming from. In general, geographical information about your visitors can contribute to your customer research and marketing database. You may even be able to discern if web visitors are coming from direct competitors, and this additional information could be valuable for your competitive analysis database.
Problems with Using Client IP Addresses or Domain Names
There are a few potential problems when you use IP addresses and domain names to identify visitor activity. These problems may cause your results to be inaccurate.
34 WebTrends Marketing Lab 2
Proxy Server Buffers
A major problem with using IP addresses and domain names as identifiers frequently arises when web visitors access web sites through ISPs or from within the network of a large corporation. When this occurs, web visitors may be routed through a proxy server before getting to the content. Consequently, it appears that the web hit comes from the proxy server rather the actual visitor. For example, most AOL users go to the Internet via a proxy server and show up as that proxy server in the reports (instead of as the actual user’s IP address). You can also have problems with aliasing across a single session when a service provider load balances using multiple proxy servers. The first hit by the visitor may be handled by one proxy server, while the next hit from the same visitor may be handled by a different proxy server to distribute the workload. When this happens, the IP address or domain name of the proxy server gets logged, making it appear that the hits came from separate visitors. Those visitors are the proxy servers, however, not the actual client machine.
Similar to the problems mentioned in the cookies section, when multiple users visit your site from the same machine, or when a single user visits your site from more than one computer, associating visitors to web activity via a computer’s IP address cannot be done accurately.
Combining IP Address and Agent Information
The next best method by which to identify unique visitors involves the use of IP addresses in combination with agent information (the client’s browser and platform). For more information, see “Collecting Web Activity Data” on page 21. IP address and agent information allow you to get around the problem of multiple visitors who use the same IP address through proxy servers, because on each machine behind any given proxy, each visitor often uses a different version of the browser. Therefore, you can get a clearer picture of the visitor based on the browser and platform used.
Probably one of the most commonly used and most accurate methods of tracking visitor sessions is through the use of a persistent cookie. A cookie refers to some text that a web server sends back to a client machine the first time that client machine visits a web site. This cookie text gets stored on the client machine’s hard drive, and in subsequent requests to that web site by the client machine, the cookie is sent to the web server. Here’s an example of a typical cookie text:
The following graphic shows the cookie process:
Here’s the process in three steps: 1. The client machine sends a request to the web server of a particular site for the first time. At this point, the client machine has no cookie information for that web site stored on its hard drive. 2. The web server processes that request and recognizes that the client request contains no cookie information. It then serves up the content requested by the client machine plus a cookie. Of course, for the cookie to function as a visitor ID, the cookie text delivered to the client machine must be unique. The web server also specifies a domain for which that cookie is valid. This way, the client machine knows which cookie to send for a given site since client machines may have hundreds of cookies for a variety of web sites. 3. The cookie gets stored on the client machine’s hard drive, and during subsequent visits to the web site, the client sends the cookie to the server in the request. The cookie is logged into the cookie field of the web server log, and may be used later to associate the visitor to all other logged hits containing that same ID in the cookie field. SmartSource Data Collector (SDC) has a cookie server component that delivers a cookie to a visitor if that visitor is new. Subsequent visits by that same visitor result in the cookie, which contains the visitor identifier, being sent to SDC along with the web activity information. The cookie is generated by SDC and consists of the IP address sent in the original request appended to a decimal-separated number based on the time stamp of the request. Because the decimal-separated number uses the time stamp down to the nanosecond level, this combination results in a number that is almost guaranteed to be unique.
Persistent Versus Session Cookies
You can issue two types of cookies: persistent and session. A persistent cookie is one that is written to the disk on the client’s computer. Therefore, it can stay or “exist” for an extended period of time. A session cookie is never written to the disk of the client’s computer. It “exists” for the length of the session and expires at the end of the session or when the visitor’s browser is closed. Therefore, the session cookie “lives” only in memory and for the duration of the session. Most companies and organizations prefer to use persistent cookies. Nevertheless, session cookies are useful, because they allow web servers to track visitors throughout a session. The federal government, for example, uses session cookies so that it does not put data on a client’s computer, because this is a privacy issue.
36 WebTrends Marketing Lab 2
If you use persistent cookies, WebTrends can recognize visitors over a period of days or longer. If you use session cookies, WebTrends can still recognize visitors who are coming via proxy servers or are sharing IP addresses, because the session cookie provides a unique identifier for that session. Accuracy regarding correlation of behavior within a visit is very accurate, but the unique visitor count will be too high because every visit will be seen as from a unique visitor.
Problems with Using Cookies
There are a few potential problems with using the cookie field to identify a visitor’s activity. These problems may prevent you from using cookies at all or cause your results to be inaccurate.
Users Who Share Computers
Consider a situation in which a family has one computer at home. Let’s say that Dad goes to a home improvement site to look at patio furniture. Let’s also assume that this is the first time anyone has used that computer to visit that home improvement site. Later on, Mom goes to that same site to check the store hours. Because a cookie was created when Dad first visited the site, when Mom visits that site, the cookie generated by Dad’s visit is sent with her request. When the web server logs are analyzed, it erroneously appears that Mom and Dad are the same visitor. Web sites also identify people who share login IDs as the same visitor.
Users with More than One Computer
Bill Smith works in a cubicle for a high-tech company and occasionally surfs the web while taking a break. He’s been interested in purchasing a bicycle lately, and for the last few weeks during his breaks he has been researching several different bicycle models. He finally figures out which one he wants to buy, so when he gets home, he jumps right to the shopping cart portion of the site and immediately makes a purchase without conducting further research. An analysis of his activity would be inaccurate, because all his research would be tied to his work computer’s cookie for that site, while his purchasing behavior would be tied to his home machine's cookie for that site. Instead of making it appear that he was a visitor who conducted a fair amount of research and then made a purchase, it would seem that two people visited the site: one who did a lot of research and then did not make a purchase, and another who did no research, but immediately made a purchase.
Cookies that Expire or Are Deleted
Sometimes web visitors decide to delete their cookies or—for a number of other reasons— their cookies get deleted. When this happens, any previous site activity associated with that deleted ID cannot be related to any new activity on the site carried out by the same person. This can also happen if a cookie expires before a user returns to the site.
First-Party Cookies vs. Third-Party Cookies
If you have a web server on your site and it serves cookies directly to your customers, these cookies are called first-party cookies. If your web page contains resources that are on a different domain and these domains serve cookies when their resources are accessed by your visitors, these cookies are called third-party cookies. Since WebTrends uses first-party cookies by default, it avoids the problems inherent with third-party cookies. As a best practice, you should use first-party cookies to identify unique visitors.
Negative Impact of Third-Party Cookies
Analytics solutions that rely on third-party cookies to identify unique visitors cause a number of business issues when these third-party cookies are rejected or deleted by a visitor.
Visitor Identification 37
Inaccurate Visitor Metrics At its most fundamental level, if an Internet user has configured the browser security settings to automatically reject third-party cookies, that visitor will not be properly counted in your web analytics results. Market estimates project this to be anywhere from 12% to 28% of Internet users on average.
Deceiving Retention Based Metrics Taking this one step further, if "John Doe", visits your web site on May 15 and accepts the third-party cookie, he will be recognized as a new visitor. If John then deletes all of his third-party cookies with his anti-spyware application on May 16th and returns to the site on May 17th, the analytics solution will identify John as a new visitor, since he no longer has the cookie on his computer. This would have an impact of under representing your retention-based metrics such as repeat visitor rate.
Inaccurate Conversion Metrics Cookie deletion also has an impact on your conversion rate for new visitors versus repeat visitors. Conversion rate = (Action taken/New (or repeat) visitors) X 100 As described in “Deceiving Retention Based Metrics,” if the cookie is systematically deleted, repeat visitor rates are undercounted and new visitor rates are over-counted, skewing the conversion rate metric you use to guage your site’s overall effectiveness.
Unreliable Campaign, Search, and Merchandising Reports In addition to tracking the behavior of a visitor to the site in general, many analytics providers correlate visitor response and site interaction to a specific campaign, search engine ,or product in an attempt to understand precisely which campaign or merchandising offer prompted the Internet visitor to take an action; much of this information can rely on information stored in the cookie. If the cookie is rejected or deleted from the Internet visitor’s browser, reports designed to identify latent or deferred conversion to a campaign or merchandising offer will be misrepresented. Also note that the longer that you track conversion to an individual marketing activity, the more likely it is that your metrics are inaccurate, as the likelihood the visitor deletes the third-party cookie increases.
WebTrends First-Party Cookie Solution
By issuing first-party cookies from WebTrends, your benefits include: • • • More accurate web results, reducing third-party rejection and deletion rates Compatibility with data collected from existing WebTrends data sources that previously used WebTrends third-party cookies Continued use of WebTrends Analytics On Demand account without needing to retrain your WebTrends users
WebTrends provides two ways for you to set up your first-party cookie solution. Each method is independent of the other. You only need to choose one method for your first-party cookie solution: • • WebTrends uses the cookie that your web server sends to visitors. WebTrends helps you serve a first-party cookie through a cookie script or cookie plug-in that WebTrends can read.
You can use WebTrends Administration to specify your first-party method and configure first-party cookie tracking.
38 WebTrends Marketing Lab 2
Session IDs or IDs Embedded in URLs
Certain web sites, especially those with shopping cart pages and registration pages, insert a unique web visitor ID into the URL. This ID is recorded in the web data activity file as part of the URL field. WebTrends can use this ID to identify the web visitor by stripping it out of the URL, and then including it in the cookie field of your web data activity file. WebTrends can then use the cookie field to sessionize your hits. The major restriction to using this method is that every web page URL for the site must contain the unique visitor ID; otherwise, visits to pages without the visitor ID will appear to be from a new visitor. Here is an example of a visitor ID in the URL field:
Some web sites attach a session ID to the user’s activity, and this ID is either recorded directly to the cookie field or in the URL query parameters of the web data activity file. Similar to processing visitor IDs, WebTrends can cut the session ID out of the query parameters field and include it in the cookie field, but session IDs—as the name implies—are only good for a given session. They do not persist across multiple sessions and cannot be used to track unique visitors.
Authenticated User Name
One of the most accurate way to identify visitors is by using the authenticated user name that visitors provide to access restricted portions of a site. After visitors logs in, you have a very reliable method for tracking their actions. However, you also need to use authenticated user names with another strong identification method such as first-party cookies to tie visitors to their referrers. Both web servers and web anaytics tools that use client side tagging can capture authenticated user name information. This would an extremely reliable method if a web site made its entire site password protected. However, there are many reasons that web sites tend only to password-protect portions of a site. Typically, these are areas of content that the visitor paid a subscription to access, as in the case of an online newspaper, or pages in which the user enters information that they wish to keep secure, such as credit card numbers, contact information, and other personal data. To identify every user as an authenticated user, the entire site would need to be password-protected, so that each visited page would result in the user name being logged in the authuser field. Consider the Yahoo sub-site, My Yahoo. To gain entrance to My Yahoo, you first have to register for the site. your first name and last name, your address, your email address, your phone number, your zip code, and perhaps answered a survey with information about your background such as single versus married, income level, interests, occupation, and more. Yahoo takes the registration information that you entered and creates an external visitor database. Each time you log in to the site, you enter your username and password. That user name shows up in the authuser field for any web data activity file hit made to an authenticated area of the site. The value in the authuser field is then used as a key to tie these hits to your visitor characteristics data in the external database. Therefore, anytime a visitor visits a site, no matter what computer that visitor logs in from, the user name remains the same. By using authenticated user names you can also eliminate aliasing that occurs when two or more visitors use the same computer to access a site. Each user must enter a unique user name and password.
In order to gain more meaningful insight into visitors’ behavior on your web site, you need to be able to assign each page view to the visitor responsible for that page view. You then need to be able to look at a specific visitor’s activity and determine that this activity occurred during one continuous visit session or over multiple visit sessions.
Visitor Identification 39
The key to all this is how you associate a visitor with each page view. There are several different identifiers that you may use to do this: • • • • • • Client IP address or domain name Combination of IP address and user agent information Cookie (persistent or session-only) Session IDs Data embedded in the URL Authenticated user name
A cookie, session ID, or authenticated user name provide very strong visitor identification, though you will probably have to do some background work to use these as identifiers. Your other common options are an IP address or IP address with user agent information. These two identifiers are readily available, but both are severely limited in how accurately they can identify visitors. Determining how your visitors behave on your web site is one of the most powerful aspects of web analytics. For this reason, considering investing the time that it takes to employ one of the more accurate means of identifying your web visitors.
Finding the Features in WebTrends Products
You will find the topics discussed in this chapter in WebTrends Analytics. Paths to the features in WebTrends Analytics:
Session Termination Time Frame
In the left pane, click Administration > Web Analysis > Options > Session Tracking. Click New.
In the left pane, click Administration > Web Analysis > Options > Analysis. Click Domains.
IP Addresses, Cookies, and Authenticated Usernames
In the left pane, click Administration > Web Analysis > Options > Session Tracking. Mouse over a session tracking definition and click Edit on the Action menu.
40 WebTrends Marketing Lab 2
Visitor Identification Worksheet
Use the following worksheet to determine who your visitor really is.
How accurately do you need to identify unique visitors? Do you assign cookies to clients who visit your site? If so, do you use persistent cookies? Or do you use session cookies? Do you want a hosted service to handle all of the cookie information? Do you want to require authenticated user names from your visitors? Do you want to keep an authenticated user database and manage it? If you have to migrate users from one system to another, are you prepared to migrate the database that contains authenticated user names? Do you want to use DNS? Will DNS slow down your system too much? Factor the time cost.
42 WebTrends Marketing Lab 2
After you understand how to collect activity data and what it looks like (Chapter 3), and you understand the concepts involved in identifying your visitors (Chapter 4), you are ready to understand how to convert this raw activity data into something that matches the organization of your web site. WebTrends web analytics provides a set of preconfigured reports on a variety of visitor behaviors—the top pages visited, the top visitors, the top entry pages, the top referrers—all standard information available from data files whether captured traditionally or using a client-side tag. The following graphic shows a sample Pages report.
To create basic measurement reports, you don’t have to do much more than tell WebTrends where the web activity data is located. Basic reports can be useful indicators of general web site activity, but there’s a lot more you can learn from WebTrends if you’re willing to put in a little effort. The real benefits of WebTrends are found when you use it to identify and improve those areas of your site that are not working optimally or are reflecting traffic patterns far different than what you expected. For example, are people linking to a specific page on your site after viewing an advertisement that you intended for them? If not, you may want to reconsider the advertisement. Do people who begin to make a web-based purchase actually complete that purchase? If they abandon the purchasing process, then perhaps it’s time for you to examine that process more closely. So how can you determine whether your web site provides the functionality and gets the results that you intended? The answer is by understanding how your site is designed and then focusing your web site analysis on those functional site areas. Specifically, you need to tell WebTrends what the specific parts of your site were created to do.
Focusing the Scope of Analysis
It can be overwhelming to try to figure out what’s happening with every single page of a large web site. Most people within an organization have an interest in specific areas of the site, not the entire thing. For example, if you work for a large company that sells computer processors to consumers and businesses, but your focus is on consumer sales, your primary interest is tracking content that is related to consumer sales (unless of course, you were comparing consumer sales versus business sales). In other words, you need to focus on analyzing the parts of the web site that matter to you.
So how do you focus your analysis on just the web site content that matters to you (or to the person who asked you to report on this content)? The answer is actually straightforward: tell WebTrends which pages, groups of pages, and other web-based content you want to examine. In WebTrends terminology, this is referred to as URL classification. URL means Uniform Resource Locator. The URL is the address of a resource, or file, available on the Internet. The URL contains the protocol required to access the resource (for example, http or ftp), a domain name that identifies a specific computer on the Internet, a directory and file path on the computer, and often query parameters used for dynamic web sites. This section discusses query parameters in more detail later in this section. The following graphic shows the URL format.
If the URL is the address of a static web page, then query parameters are not required, because the content and format of the page is the same for every visitor who requests it. For example, a page on the Internet may be located at http://www.ietf.org/rfc/rfc2396.htm. This information describes a web page to be accessed with an HTTP (web browser) application that is located on a computer named www.ietf.org. The pathname for the specific file in that computer is /rfc/rfc2396.htm. If the URL is the address of a dynamic web page, then query parameters are included with the page request. These parameters, not the page names, identify the page’s content. The dynamic web page is a way to dynamically generate larger sites from database architecture, making it significantly easier to maintain pages as the site grows.
44 WebTrends Marketing Lab 2
For example, http://clothingshopping.com/category.aspx?catID=211 indicates a specific page at clothingshopping.com that sells children’s clothing. In URL classification, you use a page’s URL and perhaps also its URL query parameters to identify and then classify that page according to its function.
Example of URL Classification
For example, on a product’s ordering page for a site that sells phone accessories (for example, Zedesco Communications), a visitor could select a cell phone cover from the products list, and then select sunburst yellow for the color option. The URL that would appear on the page might be www.zedesco.com/cart/order.asp. To learn which product is being selected, however, you need to examine the URL query parameters. In the example of the sunburst yellow cell phone cover, the URL, followed by the query parameters could look something like this:
www.zedesco.com/cart/order.asp?order_ID=10334& product=cellaccessories&type=cellcover&opt_type=color&opt= sunburst%20yellow
You could classify the page using only the URL stem (cart/order.asp) to collect all visits to the order page, regardless of what type of product was ordered. In this case, the function of the pages would be to let web visitors order products. However, to get more information, you would use the URL query parameters to classify the page visit in more detail. In this case, you would classify the page as belonging to the group of cell phone accessories items ordered. WebTrends Analytics solutions allow you to easily associate URL query parameters with an item or a group of items ordered.
This book draws on examples from a hypothetical company called Zedesco Communications that sells electronics. Consequently, this book often refers to the Zedesco Communications web site, www.zedesco.com.
URL Classification and the SmartSource Data Collector
Although the concept of URL classification was developed for web server log entries, the WebTrends SmartSource Data Collector (SDC), which collects web activity data using the client-side tagging method, also relies on URL classification to track specific pages. The way it goes about doing this, however, differs from the method used by web server logs. Instead of waiting to perform URL classification on web data activity files after they have been created, SDC applies URL classification as the web data activity file is being created, which increases performance and efficiency of the data collection process.
WebTrends Methods of URL Classification
WebTrends offers several different types of URL classification, with each method designed to help track a specific function. Some of the types of URL classification available include: • • • • Content groups Product groups Scenario analysis Advertising views
Defining Behaviors 45
Content groups designate pages with related subject matter. This grouping allows you to track the visitor interest in subject matter rather than in individual pages, which makes interpreting visitor interest far more intuitive. By grouping together related pages, you can also track web activity on your site from perspectives that may not be inherently possible with your site’s current organization. Let’s look at two example of content groups: one for a site with static web pages and another for a site with dynamic web pages.
Content Group Example (Static Site)
On a web portal that contains information such as stock quotes, news articles, and weather, you may wish to compare visitor interest in domestic versus international news. To do this, you might create a content group called “international news,” which contains all international news articles, and a content group called “domestic news,” which contains all domestic news articles. If the content is posted on a static site, you would likely have a structure of
news/international/article1.htm news/international/article2.htm news/international/article3.htm
news/domestic/article1.htm news/domestic/article2.htm news/domestic/article3.htm
These content groups specify that you gather visits to some pages in the international folder and visits to other pages in the domestic folder.
Content Group Example (Dynamic Site)
A dynamic version of this site would require that you use the parameters of the requested URL to group each related article in the right content group. A visit to an international and a domestic article on such a site might appear as:
default.asp?div=news&type=international&article=1 default.asp?div=news&type=international&article=2 default.asp?div=news&type=international&article=3
default.asp?div=news&type=domestic&article=1 default.asp?div=news&type=domestic&article=2 default.asp?div=news&type=domestic&article=3
In this case, you would track the page default.asp that had the parameter div with a value of news and the parameter type with a value of domestic or international. With web server logs, you tell WebTrends which pages belong in each content group. As WebTrends processes the records, it looks for entries that belong to a given content group. By contrast, when using a data collection server, content group information is accumulated as the pages are served. This is because when pages are created, if they belong in a specific content group, you can include the name of the content group in the page’s META tag information. SmartSource Data Collector (SDC) knows to look for this information, and then sends it on to WebTrends for reports or to a web data warehouse. By using SDC, you only have to configure a page one time to associate it with a content group. Of course, even if your are using SDC, you can still configure WebTrends to recognize content groups from the raw URLs.
46 WebTrends Marketing Lab 2
The following graphic shows a sample Content Groups report. This report identifies the most popular groups of web site pages and how often they were visited.
Product groups are a specialized type of content group that help you to track pages specifically related to products you sell or promote on your site. WebTrends analysis products track product groups separately because products are such a high profile component of most sites.
Product Group Example
Let’s say that Zedesco wants to track web activity visits to content about cell phones and cell phone accessories. To do this, they create a product group that includes product pages for cell phones and their accessories. If one directory contains all the cell phone content and no other type of content, they can simply specify that directory. However, if cell phones and their accessories are stored in different directories, and other, non-cell phone content is included as well, they will have to do a little more work to define their product group. Assume that the site is structured with the following product pages:
products/phones/cordless phones/SBC-2905.htm products/phones/cordless phones/SBC-7205.htm products/phones/cordless phones/SBC-3205.htm products/phones/cell phones/XT2100.htm products/phones/cell phones/SCH-N300.htm
products/phones/cell phones/N-3285.htm products/phones/accessories/travel-charger.htm products/phones/accessories/covers.htm products/phones/accessories/headset.htm products/phones/accessories/videogame.htm
Keep in mind that some of these pages represent cordless phones and others represent adjunct phones.
A large, database-driven site that uses dynamic URLs would use the following structure: products/info.asp?prod=1783&cat=13 where 13 represents cordless phones 1783 identifies SBC-2905 If you wanted to report on all phones in a product group, you would capture the following, assuming that the travel chargers, car-kits, headsets, and the video games are cell phone accessories:
products/telephones/cordless/caller ID/VT910ADL.htm products/telephones/cordless/caller ID/VT9125.htm products/telephones/cordless/answering system/VT910ADL.htm products/telephones/adjunct phones/corded/panasonicKXTMC97B.htm products/telephones/adjunct phones/corded/GE29870GE1.htm products/telephones/adjunct phones/answering systems/GE29992.htm
However, note that in this example, the cordless phone VT910ADL can be categorized with both caller ID capable systems and those with answering systems. It is common for pages to have several places where they might be logically grouped. To categorize phones, you configure WebTrends to take all content in the \products\telephones\ directory, and group them with the individual pages for the remaining items. In this case, you would configure WebTrends to group visits to the cordless phones/caller ID directory and the cordless phones/answering systems directory and track visits to the pages in those directories. The following graphic shows a sample Product report. It represents the number of visits during which product-related pages were viewed.
48 WebTrends Marketing Lab 2
In the context of defining your site’s structure for WebTrends, you need to know which areas of your site, if any, contain sequences of pages that make up a web-based task you want your visitors to complete. These sequences of pages are called scenarios. Some of the most common examples of scenarios are registering as a user of a web site, making an online purchase, or filling out a survey. For example, Zedesco has a purchase process that requires web visitors to fill out the following pages to complete their purchase: • • • • Product page viewed. Product added to cart. Checkout started. Checkout completed.
These steps constitute a purchase scenario. Other, less familiar sequences on your site may also be important to track. For example, a sequence of product pages that you want to make sure visitors are viewing, or if you are a travel web site, a set of pages that your visitors must complete to track prices for their top flight itineraries.
The following graphic shows a Purchase Conversion Funnel report. This analysis offers insight into each step along the information request process. Each step shows a drop-off as visitors move through the funnel.
50 WebTrends Marketing Lab 2
If your company hosts advertisements on its site, it can be very important to show your customers how much traffic the ad you’re hosting for them generates. In addition, the development of pricing schedules may be heavily dependent on where the ad is placed. You may need to provide numbers to potential customers that show how valuable a particular piece of web real estate is for advertising. Reports on traffic generated by ads placed in various areas of the site can let your customers balance level of exposure versus cost when making their decision about posting their ad. Advertisements can be broken into two parts: • • Ad View – Visitor views a page containing the ad graphic or link. Ad Click – Visitor actually clicks on the ad and opens its content.
Depending on the ad hosting method, both the ad itself and the content it links to may be hosted on your site. However, it is also common to host the ad on your site, yet have the content of that ad hosted by your customer, on their site. In the first method, the Ad View and the Ad Click that results in the ad content display are both logged to your web server log because all activity occurs on your web site. In the second method, the Ad View activity is logged to your web server log, but the act of displaying the ad content display is logged to your customer’s web server log, not yours. You can get around this issue by implementing server-side scripting (for example, CGI, Perl, or ASP) to perform a redirect to the destination URL. A very common Perl script is redir.pl. This redirect command sends the hit information back to your web server’s data activity file, and is recognized as an indicator that the ad was opened. Of course, if you are using a data collection server or client-side tagging method, you can easily collect this information by running a script each time an Ad View or Ad Click occurs. An ad click is an indicator of greater interest in the ad than an ad view is because it implies that the user focused directly on the ad and was interested enough to click on it.
The following graphic shows an Onsite Ad Impressions report that shows how often specific ads were viewed.
In the Onsite Ad Impressions report note that the Ad Views Visits column refers to the number of visits by visitors who saw the specified ad. A visit is a series of actions that begins when a visitor views a first page from the server and ends when the visitor leaves the site or remains idle beyond the idle-time limit. The default idle-time limit is thirty minutes. This time limit can be changed by the system administrator. Therefore, a visitor may see an ad more than once during a visit, but the ad will only be counted once in this table and graph.
Other Site Structure Issues
Before you begin to tell WebTrends how your site is structured by the patterns it can find in your URLs, you need to define your home page definition and how to classify files based on their extensions. These two items affect how WebTrends counts visits to your site’s home page and how it interprets the files you request. Unless your home page name changes, typically you specify these settings just once; whereas the settings you configure in URL classification are likely to change according to your analysis needs and according to modifications of your site.
52 WebTrends Marketing Lab 2
Handling Dynamic Pages—URL Rebuilding
Sites that make heavy use of dynamic pages require a little extra thought. Typically, dynamic sites are driven from a few scripts that use parameters to control the content of each page as it appears to the visitor. Yet the name of the script by itself—for example, default.asp or catalog.php—is not very descriptive of what the visitor sees. These names do not illuminate reports that list the top pages or show paths through a site, because it looks like only a handful of pages are visited. For example, listing pages only by the URL file name could result in the following statistics: Page
Parameters used to request those pages control their actual content, and so it is those parameters that need to be included along with the page name. For example, using the dynamic URL:
You may find it most informative to know which division and type of articles are being viewed. It makes sense to include those parameters in the page’s URL for reporting. You should not include parameters that are unique for each visitor session, such as the session ID, because doing so makes every page request appear to be different content. WebTrends allows you to “rebuild” the URL, specifying which parameters to use. In the example above, you may want to include the “div” and “type” parameters only. This could be used to transform the URL above into:
Using the URL rebuilding feature, the Pages report becomes more enlightening: Page Visit s
528 431 366 132 89 67 44
default.asp?dif=news&type=domestic default.asp?dif=financial&type=domestic default.asp?dif=ads&type=domestic default.asp?dif=news&type=international catalog.php?dept=clothing catalog.php?dept=hardware catalog.php?dept=kitchen
6,243 2,511 3973 674 694 185 56
Note that the parameters are sorted alphabetically. This ensures that two URLs which differ only in order or parameters are still considered to refer to the same content.
Home Page Definition
Counting visits to your home page can help you determine whether the bulk of your visitors come to your site via your home page, or whether they entered your site from somewhere else —perhaps from a bookmarked page, an ad, or some other link. The home page, just like any other page, has a file name. But typically, especially when visiting a site for the first time, the visitor types in the site’s name. Why? Because most sites names are easy to remember. The visitor may even know the division within the site, so will enter a folder such as “products” or “news” after the site’s name. The file name for the home page—whether it is the top-level home page or the home page for a division within the site—is not so easy to remember. Consider the variety of possible names that you see: default.htm, default.asp, index.htm, index.asp are all standard names, but there’s nothing that prevents a site designer from making up a completely unusual name. For this reason, web servers are designed to recognize that the visitor is attempting to visit the home page when they omit the home page’s file name, and serve up that content. If a visitor entered www.zedesco.com, the web server would deliver the home page, www.zedesco.com/default.asp. The problem is that whatever is requested gets recorded in the web data activity file. You want the entries to be viewed as the site’s home page, not separate pages. To make this happen, you have to tell WebTrends that web data activity file entries that appear as: GET/ and GET/default.asp are actually visits to the same page—the home page. This allows you to obtain an accurate count of home page visits.
As web site development and publishing have become more involved, so have the types of content that can be hosted by your site. In addition to standard HTML documents, sites also host downloadable files for Flash presentations, Microsoft Word documents, Adobe Acrobat PDF files, compressed files, video files, audio files, executables, and so forth. You need to tell WebTrends how you want it to view various file types based on their file extensions. While at first it may seem obvious which file types are documents and which are downloadable files, consider how you might classify the following Adobe PDF file.
Whether you consider it a downloadable file or a document depends on how you expect visitors to use it. For ambiguous cases such as these, you must configure WebTrends to correctly identify each file extension. That way, when WebTrends processes the web data activity file and encounters a record such as the request for the Nokia C23 Owner’s Manual, it knows whether to count the file as a download or a document request.
54 WebTrends Marketing Lab 2
You categorize files in order to determine whether or not your visitors look at certain types of files. If you devote a substantial portion of the budget to creating multimedia pieces for your site, you want to know that your investment is paying off. You may also have the same information presented in multiple formats and want to know which format your visitors use the most: static documents or interactive elements.
Many people who use WebTrends never realize the full potential that lies in the features it provides. Instead, they only venture as far as using the standard reports that ship with WebTrends and track information about the entire site, not specific pages or areas of the site. The real value in web analytics is in identifying and examining specific areas of your site in detail. Typically, these areas are ones that allow web visitors to complete an action, such as making a purchase, researching a product, or solving an issue by reviewing online support materials. The tools provided with WebTrends allow you to track visitor behavior: visits to content and product groups, the steps in a scenario, clicks on advertisements, and the paths that visitors took through your site. All of these tools can help you focus on your site to find what is working and what needs some improvement.
Finding the Features in WebTrends Products
Click Web Analysis > Report Configuration > URL Parameters also Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > URL Parameter Analysis
Content Groups (and Product Groups)
Click Web Analysis > Report Configuration > Content Groups or Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Content Groups
Click Web Analysis > Report Configuration > Scenario Analysis or Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Scenario Analysis
Advertising Views (and Clicks)
Click Web Analysis > Report Configuration > Onsite Advertising or Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Onsite Advertising
Home Page Definition
Click Web Analysis > Reports & Profiles > Edit a profile > Analysis > Home
Click Web Analysis > Options > Analysis > Page File Types
Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > URL Rebuilding
Defining Behaviors Worksheet
Use the following worksheet to focus on your web site functionality – what is working and what needs improvement. Consideration
Your web site is organized so that it can be searched according to content groups? You need to know the visits and hits for each content group? Your web site is organized so that it can be searched according to product groups? You need to know the visits and hits for each product? Your web site has scenarios that you would like to analyze (for example, shopping cart)? You need to know the number of Ad Views and Ad Clicks on your ads? Your home page name changes now and then?
56 WebTrends Marketing Lab 2
Filtering and Analyzing Your Data
If you were packing for vacation, you wouldn’t pack all the clothes that you own. If you did, you would take much more than you need with you. Instead, you might put all your clothes out on a bed, examine what you have, and then select and pack what you need. In this situation, you would include only those items you know that you’ll use. Conversely, you might return items you don’t want to the closet or dresser drawers, leaving only the items that you do want to pack. By going through this process, you’ve narrowed down all your clothes to just the clothing that you know you will need. This not only reduces how much clothing you have to store in your suitcase, but it also saves you from having to sift through all your clothing each time you get dressed. Consider approaching your web server data files in much the same way you would when packing for a vacation. If WebTrends had to sift through all your data, your system would be working harder than it needs and would unnecessarily be using up storage space. In addition, once analysis is done, you would prefer to review results that have meaning for you, not all possible results. Filtering is the process of preparing to run a web activity analysis that allows you to select only pertinent data. Filters allow you to determine such things as new versus returning visitors, which visits were initiated by a campaign, and which visitors were internal employees or external visitors. If you choose to apply no filters to your web-activity files, the analysis software analyzes all the data. However, this may impact performance and analysis time, because your data records will contain information about images and other kinds of data that contain no real value. Keep in mind that you can also apply queries and searches to reports in WebTrends Analytics Reports. For more information on queries and searches, see the Analytics Reports Help.
Profile-Level and Custom Report Filters
First, you need to determine if you want to filter all of your activity data or just some parts of it. If you apply filters to all of your activity data, then you must keep in mind they affect all of the analysis. In most cases, you will want to filter out images (such as JPEGs and GIFs), spiders, robots, and anyone from your company who is testing the site. Keep in mind that global filters select a portion of data for analysis. You can achieve most of your filtering goals at the custom report level. With custom report filters, filters are applied to a specific report to produce reports that are specifically tailored to your needs. Through custom reports you can achieve greater visitor segmentation and site segmentation. Site segmentation means that you can examine specific areas of your web site such as the directories that deal only with a specific type of content.
Include and Exclude Filters
After you decide whether to apply filters to all your profile data or at the custom report level, you should consider whether to use include filters or exclude filters. • • Include filters specify the data to use in the analysis. Exclude filters specify the data NOT to include in the analysis.
Sometimes it doesn’t matter which filter you use, but at other times, one kind of filter is distinctly more convenient to use than the other. You can easily apply the concepts of including versus excluding data with two different levels of filtering: filtering on hits and filtering on visits. The remainder of this chapter describes how include and exclude filters work with hit filters and visit filters. By understanding the concepts involved, you will analyze data that pertains to your needs.
How Filters are Processed
For detailed technical information about how WebTrends processes filters, see WebTrends Administration User’s Guide.
Setting Up Your Profile—Initial Filtering
A profile is a group of settings with which you identify the visitor activity data to be collected, filtered, analyzed, and displayed in your WebTrends reports. Typically a profile is created for an individual web site, but often a separate profile will also be created to report on a portion of a site, or to roll multiple sites together. Through the profile, you define the data source location, any activity you want filtered from the reports, and user rights to the resulting reports. More information about profiles (especially parent-child profiles) is presented in “Parent-Child Profiles” on page 69. When you first use WebTrends, you most likely will set up your profiles, which gets the web data set up and the reports that you want to create. Then you will run your profiles. Afterwards, for deeper inspection and further analysis, you go back and filter the data in various ways.
Hit and Visit Filters
To understand why you need hit and visit filters, you must first understand the concepts of hits and visits.
The concept of hits and visits was introduced in “Defining Web Activity” on page 31.
When your web server or data collection server records visitor activity, each line in the record represents a hit to the server. Hits are the individual activities that combine to make up a visit to a single page. Think of the contents of a typical web page. Most consist of some text and one or more graphics. When users request a page, they are actually making requests for each item on the page–maybe a GIF image of a company logo, some HTML text, and a JPEG image. The server either successfully or unsuccessfully handles each item, and then logs the results of the request for that item, or hit, along with other information about the hit. One record in the web activity data file equals one hit. Actually, with web server data files, this one record does equal one hit. However, for client-server tagging, WebTrends SmartSource Data Collector server data files do not typically record hits to graphics images. In the case of a SmartSource Data Collector server log, you will typically only have page hits.
58 WebTrends Marketing Lab 2
A visit, or a visitor session, includes all the pages a unique visitor requests during a period of continuous activity on your site. Consequently, it includes all the hits associated with those pages in the visit. Visits are considered closed after the visitor remains inactive for a specified period of time. As a general rule, a visitor session should be closed if the user remains inactive for 30 minutes, although your WebTrends administrator may wish to specify a timeout period that is more in keeping with your analysis requirements.
Hit Filter Criteria
When you filter on hits, you filter in or out each individual piece based on some specified criteria, not all of them at once. Hit filters allow you to filter at a more granular level than you can with visit filters. With a visit filter, you are filtering all hits associated with an entire visit session – not so with hit filters. The following subsections discuss the criteria you can use for filtering hits. These criteria include: • • • • • • • • • • • • • URL HTTP Method Cookie Multi-homed Domain Client Browser Return Codes IP Address File Directory Onsite Advertising Day of the Week Hour of Day Authenticated Username
You may decide that you need to include or exclude certain pages from analysis so that you can focus more directly on specific areas of the site. For example, if you are part of an IT organization, you may wish to determine whether your web visitors are viewing your knowledge base articles, all of which have a prefix of “kb_”. You could either list all of the knowledge base articles you wish to track, or, since WebTrends supports wildcard usage, you could specify that your filter includes all files beginning with “kb_”. If your site uses a content management system, then instead of specifying pages to include or exclude, you may need to specify a page and any URL query parameters that grabbed the content displayed in that page. An example of knowledge base articles that you may wish to track web activity for could be for issues with the P100 cellphone. The excerpt below is a hypothetical web data activity file entry that shows how this could appear:
2001-03-04 00:25:51 proxy1.thegrid.com - W3SVC3 web1 192.168.1.1 GET /support/ default.asp product=p100&id=kb_5
The query parameters are product and id, where product=P100, and id=kb_5. You could track activity for P100 articles by specifying that your analysis include all hits with the page, default.asp, the product query parameter having a value of P100, and any records with an id value that contains the prefix kb_.
Filtering and Analyzing Your Data
Your web server log may show requests using several different HTTP methods, but most frequently, you will encounter GET requests. These requests, when logged, contain more useful information for analysis purposes than any other method. A GET request returns whatever information is identified by the request URL and associated query parameters. For example, if you are using the Internet, and you click on an image, the actual request for that image might look like this:
GET /picture.jpg HTTP/1.1
In a distant second place is the POST method, which some web sites use to post forms. A couple of other rarely used methods are PUT and HEAD. These methods seldom contain useful information for web analytics, and because they are used infrequently, they may never appear in your web data activity file. Typically, your web traffic analysis will process GET requests, though if your site has forms that use the POST method, you may wish to track activity on those forms. WebTrends has the capability to exclude records of requests using methods you don’t want to track. Of course, you could also choose to include only those methods you do want to track and the results would be the same.
As mentioned in Chapter 4 (see “Cookies” on page 35), cookies can be a means by which WebTrends can recognize visitors. However, cookies are used to store various types of information, such as shopping cart contents, time of first visit, and number of visits. By selecting an appropriate cookie, you can investigate the behavior of a specific segment of your visitors. The cookie filter is typically used for this investigative purposes. This can be useful, for instance, if you know of visitors whose activity is not pertinent to your analysis, and you wish to exclude their activity.
If your site is spread across multiple domains on the Internet, you may want to view the activity of only one domain. You may also wish to exclude the activity of one or more domains. A multi-homed domain filter lets you specify which domain or domains to filter from the analysis. Let’s say that your company is based in the US, but its site has sub-sites in the US (www.yourcompany.com), some in France (www.yourcompany.fr), and some in Germany (www.yourcompany.de). If you only wished to view the main US site, you might wish to either exclude the French and German sites, or it might be easier to include only data from the US site in the analysis. For users of SmartSource Data Collection, the multi-homed domain filter can also be used to filter out hits from sites that may have copied pages and the SDC script included in that page (recall the discussion of client-side tagging; see “Using Client-Side Tagging” on page 21). Another use (by filtering in) of the multihomed domain filter is to identify sites that have “stolen” copyrighted material.
With all the different types of browsers available today, you may want to get a sense of the types of activity carried out from various flavors of browsers—Internet Explorer, Netscape Navigator, WAP and Palm device browsers. You may even want to know if activity originated from a robot or spider crawling your site. Your web data activity files typically contain a reference to the browser used to access content. The files also record visits from spiders and robots in the same browser and browser version field.
60 WebTrends Marketing Lab 2
If your business has a portion of its site devoted to WAP devices such as cellular phones, and you wish to examine visitor activity on only those WAP-specific areas, you could tell WebTrends to only analyze requests originating from WAP browsers. The excerpt below shows a possible web data activity file entry that would be included in analysis if you created an include filter for WAP device browsers.
2001-03-04 08:39:02 126.96.36.199 - SERVER10 WEB1 - GET /wml/products/wireless/ phones.wml - 200 0 647 543 0 80 HTTP/1.1 UP.Browser/3.1.03-NK02+UP.Link/188.8.131.52 WEBTRENDS_ID=184.108.40.206-2562687908.34229567 -
A portion of this excerpt refers to the browser and browser version number used by the client making the request:
You may also wish to compare the types of activity you experience from a specific standard HTML browser such as Netscape or Internet Explorer. Because these browsers handle HTML code slightly differently, comparing the visitor experience on one browser with another can reveal valuable information. For such a comparison, you could create an include filter for each browser of interest and then review analysis results for each browser. For example, if you find that Netscape Navigator users drop out more frequently in a shopping cart scenario than do Internet Explorer users, this may indicate that the HTML code does not appear as you had intended on browsers using Netscape Navigator. Although web designers always try to review their sites in several different versions, it's easy to miss problems with design when you have numerous pages to review or if testing is not thorough.
Return codes indicate whether or not requested content was successfully delivered, and if not, what the problem may have been. Return codes in the 200s and 300s indicate a successful content delivery, while those in the 400s and 500s indicate a failed delivery. For most web visitors, the most well-known and irritating error is the standard 404 File Not Found error. In the web activity data file, this appears as a server-to-client status entry. The following data file entry shows a successful return code of 304 (Success Not Modified) in the first data file entry, and a success return code of 200 (Success OK) in the second data file entry. Both return codes are highlighted in bold print:
2001-03-04 00:03:23 computer.attcanada.ca - W3SVC3 web1 192.168.1.1 GET /club/kb/s32/ motors.wmp - 304 0 27000 58 412 80 HTTP/1.1 Mozilla/ 4.0+(compatible;+MSIE+6.0;+Windows+NT+5.1) WEBTRENDS_ID=10.14.211.5292873123.102983222 2001-03-04 00:04:09 computer.quest.com - W3SVC3 web1 192.168.1.1 GET /dealers/ default.asp WT.sv=Web%20Server%201&WT.ti=Dealer%20Home&WT.tz=420 &WT.ul=en&WT.cd=32&WT.sr=1024x768&WT.jo=Yes&WT.js=Yes&WT.co=Yes 200 0 37211 121 389 80 HTTP/1.1 Mozilla/4.0+(compatible;+MSIE+5.0b1;+Windows+NT)
Because 400 and 500-level errors indicate potential problems with your site, you may choose to create an include filter that analyzes only the activity on failed requests. You can then determine which pages may have problems that are preventing users from accessing your content and modify those pages to resolve the problem.
What if your company just launched its web site after a major site redesign? Your company had a big launch party, and all the employees afterwards decided to look at the redesign on their own. You probably wouldn't want to include their visits in your analysis, so you could simply filter them out based on their IP addresses or your company’s domain name.
Filtering and Analyzing Your Data 61
Within each web data activity file entry is a field that indicates the computer address of the visitor. Depending on whether or not you instructed WebTrends to resolve IP addresses, this may either be an IP address or a domain name. Filtering on a visitor’s IP address or domain name allows you to include or exclude specific addresses in your analysis. You might also want to see levels of activity based on regions, country, or domain types. The web data activity file entry below with the bold highlighted entry shows a visit from a computer located in Canada, as evidenced by the .ca extension:
2001-03-04 00:03:23 computer.attcanada.ca - W3SVC3 web1 192.168.1.1 GET /club/kb/s32/ motors.wmp - 304 0 27000 58 412 80 HTTP/1.1 Mozilla/ 4.0+(compatible;+MSIE+6.0;+Windows+NT+5.1) WEBTRENDS_ID=10.14.211.5292873123.102983222
If your web site caters to educational institutions, then you would be most interested in activity originating from educational organizations. You could capture this data for analysis by creating a filter that included all educational sites based on their domain type extension of .edu. Another use of the IP filter is to filter out monitoring software, such as Keynote, which is used to maintain the health of the web site. That is, companies and organizations with extensive web sites find it beneficial to have their web site monitored by special monitoring software. Every time the monitoring probes a given web site, all of its activity will be counted unless the IP filter has been used to filter out the monitoring software.
Many hits contain requests for images that have very little meaning for you. Besides overloading your system with meaningless data to analyze, you are likely more interested in the actual pages that were opened during a visit than the images your visitors saw. You can use a specific filter to select the file types, such as GIFs, JPEGs, and other image or graphics files, that you wish to exclude or include from analysis.
Use this filter for simple file name matching. For more complex filtering, you may want to use the URL filter.
62 WebTrends Marketing Lab 2
The following graphic shows a report that identifies the accessed types of files for your site and the total number of kilobytes of data transferred for each file type. The percentage column (%) reflects the percentage of all kilobytes of data transferred for the specified file type.
If your site is structured in such a way that various directories include specific types of content—the products directory contains products content, the support directory contains all technical support content, etc.—it may be helpful to look at various areas of your site by including or excluding content based on the directory and sub-directories in which that content resides. Tell WebTrends to include the directories that contain content of interest to you, or conversely, the content you wish to exclude from analysis.
Use this filter for simple directory matching. For more complex filtering, you may want to use the URL filter.
Many sites sell advertising space as a way of bolstering their income. To be able to track ads more easily, hosted ads typically consist of a graphic on a web page that when clicked, passes the user through a redirect page. This redirect page then opens the ad’s content. For both billing purposes and to assure those companies who advertise on your site that advertising on your site works, you need to show them that visitors are viewing the pages on which their ads reside, and that those visitors are then clicking on those ads to view them. Clients will typically only want to see the activity on their ads. To do this, you need to create an include filter for the ad view and ad click for each client’s ad. The following two sample hits show first an ad view of the ad graphic, specials.gif, which is hosted on the site www.austinbusinesscomputing.com. The second shows an ad click that took the user to a redirect page, yahoo1.htm, which made it possible to track an ad hosted on Yahoo.
2007-02-07 08:12:11 nsts02-1077.sts.embratel.net.br - SERVER10 WEB1 - GET /ads/ specials.gif - 200 0 17527 587 4456 80 HTTP/1.1 Mozilla/ 4.0+(compatible;+MSIE+5.0;+Windows+98;+get2net+update) WEBTRENDS_ID=220.127.116.113218603766.52660653 http://www.austinbusinesscomputing.com/ads/networkAd.htm Filtering and Analyzing Your Data 63
2007-02-08 08:11:19 nsts02-1077.sts.embratel.net.br - SERVER10 WEB1 - GET /redirect/ yahoo1.htm - 302 0 835 436 10 80 HTTP/1.1 Mozilla/ 4.0+(compatible;+MSIE+5.0;+Windows+98;+get2net+update) -
To report on the specials ad, you would filter in only the ads/specials.gif file. To report on the ad clicks, you would filter in only the redirect/yahoo1.htm file and the return code 302.
Day of the Week and Hour of the Day
Being able to analyze your web activity for only specific days of the week, or for a certain time period during the day can be useful in a number of circumstances. For example, if your site has a weekly online newspaper that came out Wednesday, you might be interested in knowing how many visitors view your content right when it appears on the Web. Or, if you are tracking employee activity on your corporate intranet, you may prefer to only track activity during standard business hours Monday through Friday. Each hit recorded in your web data activity file has a time stamp that can be used to filter in or out specific days of the week, and specific time intervals during the day. The following graphic shows a report that provides the average activity on all pages (of a web site) for each hour of the day. The percentage column (%) reflects the percentage of visits to your site that occurred during the specified hour.
Authenticated User Name
If your site requires that your users provide a user name and password to access portions of your site, you can include or exclude hits from specific visitors based on their user names. This concept is very similar to that mentioned in the cookies section earlier, though cookies can be used to filter more than just specific users. You might use an authenticated user name filter if you found that a particular user who you do not trust is snooping around on your site. You could just as easily use an authenticated user name filter to discover if a particular prospect is exploring your site.
64 WebTrends Marketing Lab 2
Visit Filter Criteria
You can use visit filters to include or exclude all activity related to an entire visitor session. The following sections describe types of criteria you can apply to an entire visitor session. These filter criteria include: • • • Entry page Referring URL or site Advertising campaigns
Filtering visits is slightly more restrictive compared to filtering hits. With hit filtering, you apply the filter to all hits in the log file. Any hit that matches the criteria is either included or excluded from analysis depending on the type of filter you specified. When you filter on visits, however, the web data activity file has been parsed and processed to sessionize your data. At this point, you are not actually applying the filter to the raw web data activity file—you are applying it to a summary of the hits associated with a visitor session.
The page on which the visitor first enters your site is the entry page. Filtering by entry page lets you include or exclude from analysis visits that started on specific pages. For example, if you have a redirect page you’re using to track an ad, you might choose to include only activity associated with visitor sessions that began with a click through the ad’s redirect page. You might also want to view only activity of visitors who began their visitor sessions somewhere in the middle of your site, because these visitors often have more of a purpose in their visit than do visitors who enter at your home page. To do this, you would create an exclude filter that filters out all visits to your home page.
Filtering and Analyzing Your Data
The following graphic shows a report that identifies the first page viewed when a visitor visits your site, the number of visits to those pages, and the percentage of times this page was the entry page compared with other entry pages. The most common entry page is usually the home page, but other common entry pages include specific URLs that visitors type, pages that have been bookmarked, or pages referred to by other sites.
Referring URL or Site
You might also wish to exclude or include all visitor sessions that were started from a particular referring URL or site. A classic case where you might use this is if your web server log contains many self-referring visits. This happens when the visit session times out due to inactivity, but the visitor is actually still on the site. When they resume viewing content on your site, it will appear as though they have started a new visit session, even though they were already on your site. When this occurs, the referring URL that gets recorded in your web server log is the page on your site that they were last viewing before beginning the “new” visitor session.
66 WebTrends Marketing Lab 2
If you have advertising campaigns on your site you may wish to track the activity that is occurring on them. To do this you must create a campaign definition before you can filter on a campaign. This definition specifies the referring page or entry page that, when visited, represents a visit to the campaign, or, in the case of SmartSource Data Collector, query parameters. The most common use for filtering by campaigns is to include only the visitor session activity associated with a particular campaign. If you have a reasonable idea of the value that you can associate with specific activity, you may be able to forecast the revenue that can be generated by the campaign. The following graphic shows a report that provides visitor activity for each campaign.
Filtering and Analyzing Your Data
WebTrends analysis products ship with a number of pre-defined reports that cover the information most organizations want, but every organization has its own, unique requirements for the web activity information it needs to see. This is where custom reports are particularly useful. Custom reports allow you to set one or two table dimensions—for example, you might want information about new visitors from a specific geographical region or with a certain income level. With custom reports, any dimension for which you have data, including any external data source you may have tied to your web activity data, can be tied to measures such as the number of page views, the number of visits, or the duration of a visit. If you need to narrow down what you view in the reports, you can apply filters to the report data just as you did when filtering the summary tables. WebTrends provides numerous dimensions for custom reports. Here are a few examples: • • • • • • • • • • Lifetime Value Range Most Recent Campaign Most Recent Search Phrase Product Manufacturer Daily Buyers Daily Visitors Order Page of Interest Product Cart Adds Visitor Purchase Count
WebTrends also provides numerous measures for custom reports. Consider the following examples:
Not every measure-dimension combination makes sense. Some dimensions are very large and should be used wisely. For example, you don’t want to use unique visitor with referrer, because the virtually unlimited number of unique visitors and referrers would overwhelm your tables. Custom reports support data look up that translates coded information from your database into more meaningful descriptions. See “Campaign IDs and Translation Tables” on page 78. Here are descriptions of several custom reports that may be helpful when you consider the data you might want to analyze:
Customer vs Non-Customers
Lets you see how many of your web site visitors purchase products from your web site. Compare the number of visitors who make purchases (buyers) to those who do not (non-buyers) by time period.
Content Group Duration
Provides insight into which areas of the site are most attractive to your visitors. Analyze the content areas for possible cross-promotions, or analyze over time to interpret content popularity.
Shows activity occurring during the report time period segmented according to the demand channel of the last campaign to which a visitor responded.
68 WebTrends Marketing Lab 2
Provides a drilldown presentation of the geographical information (region, country, state/province, city) relating to the visitor’s IP address. The WebTrends GeoTrends Database is required to get complete information down to the state and city level.
Shows the marketing programs for the most recent campaigns that drove traffic to your site during the report time period. For the report time period, all conversions and other activities are tracked and attributed to the last campaign to which visitors responded. Thus, even if the conversion does not happen on the first visit generated by the most recent campaign, the appropriate source is “credited” with the conversion.
Purchase conversion funnel by search phrase (all)
Helps you understand how the usage of all search engines and phrases correlates to conversion activity on your site. This report includes both organic (for example, natural search) and paid (for example, pay-per-click) search referrals. The conversion funnel allows you to analyze each step of the purchasing process to determine specifically where users are dropping off and which percentage completes the checkout process.
Sales cycle by product
Shows the number of days between a new buyer’s initial visit and first purchase for each product. The following graphic shows the number of days between a new buyer's initial visit and first purchase.
Dividing the web traffic rather than filtering it is often an efficient alternative to custom reports. Many companies have one web site and one web server that generates all of their web-activity data files. In particular, large companies with many divisions may require a more complex way of dealing with their data files, because each division may have responsibility for a portion of the web site. Since each division will want reports that are tailored to the needs of that division (but not to the needs of other divisions), you have to generate hundreds of different kinds of reports. All the activity is gathered in one data file, but you don’t want to reprocess that data file hundreds of times to get the reports. You want to read the data file once and generate all of the reports that you will need for each division and a summary report. The reports will be basically the same, except that each report will contain only the specific piece of data that relates to a particular division of the company. The parent, then, is the company at large and the child is each division.
Filtering and Analyzing Your Data
What this means is that parent-child profiles/reports are typically used by multi-domain organizations (for example, service providers or large corporations) to simplify administration. A parent profile specifies the global settings that will be applied to any child profiles, and specifies when to create a child profile. In many cases, the presence of a new domain or sub-domain could trigger the creation of a child profile, or in some cases, the presence of a parameter in the URL is used. An example of this would be the creation of a child profile for a major content area of a site, if a complete set of reports is required for that content area. The parent profile automatically creates child profiles based on your criteria, which point to a limited set of your web data. The child profiles then analyze the subsets of your data. Parent-child profiles/reports can also be applied to content groups. You may be interested in the web activity for a particular content group, and you may have a number of different content groups that you want to examine. Therefore, several divisions of a large company could be interested in the reports relating to a particular content group. The parent in this case, is the company at large, but the profile/reports on a content group represent the child.
Reducing Profiles and Increasing Productivity
Generally, there are a couple of major reasons why you may want to establish new profiles: • • To see the full set of reports on a subsection of available data To apply different filters to divide your web activity data into segments.
If you would like to report on and analyze a particular portion of your site, you can create a new profile that only considers that section of your site. But if you look at the depth of analysis you need for this section of your site, creating hundreds of reports all specific to that section of the site may be overkill. It might be best to instead create a few custom reports that show you the traffic volumes and campaigns that are driving traffic to that section of the site. Likewise, if you need to apply different filters to the same segment of data (for example, one campaign versus a second campaign), you could create a separate profile for each campaign. Again, though, it may be excessive to create the hundreds of reports created by a full analysis profile. Better instead to consider custom reports for each campaign.
Filtering allows you to narrow down the volumes of web activity data to just the data you want to examine. Different types of filters can be used to focus on just the types of data you wish to analyze. You can apply filters to each line in the web activity log using hit filters, or you can apply filters to visits using visit filters. Visit filters are applied after the individual hits have been filtered on and the visit data has been sessionized. You may also specify which data to include or exclude from your reports. Indirectly, this is a way of reducing, or filtering, the data that you see in reports. The benefits for filtering data include not only reducing the amount of data that you need to store in your tables of aggregated data, but also making the amount of data you do want to examine more manageable.
Finding the Features in WebTrends Products
You will find the topics discussed in this chapter in WebTrends.
Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Hit Filters or Visitor Filters
Click Web Analysis > Report Configuration > Custom Reports
70 WebTrends Marketing Lab 2
Use the following worksheet to help understand what kind of filters you need. Consideration
Do you plan to use image files (such as .jpg, .gif, or .tif files in your analysis? Do you plan to include spiders and robots in your analysis? Do you plan to include hits from people within your own company who look at your web site? Do you need high-level reports on ad campaigns or more reports on browsers and technical information? Is visitor segmentation important to your analysis? Is site segmentation important to your analysis? Does your company have many divisions requiring parent-child profiles?
Filtering and Analyzing Your Data
72 WebTrends Marketing Lab 2
Nearly every web site shares three fundamental web analytics objectives: acquire more qualified visitors for the lowest cost, convert these visitors into customers, and retain these customers for repeat business. • Acquire more qualified visitors From online marketing to offline marketing, the first step in winning new customers today is driving new traffic to your web site. But all traffic is not equal. You need to drive the most qualified visitors for the lowest cost. With WebTrends you can get a complete picture of campaign response, campaign conversion and overall return on investment (ROI). As a result, you can pinpoint exactly which campaigns are working and which aren’t. This chapter discusses acquisition in more detail. • Convert more visitors by analyzing click-by-click behavior Whether your web site goal is for visitors to register, make a purchase, or get technical support, conversion rate is a critical measure of your site’s success. WebTrends provides the most comprehensive navigation analysis in the industry, allowing you to track visitors click-by-click, identify confusing navigation and minimize abandonment. Isolating problem areas in your site and experimenting with improvements can have a big payoff. For more information about conversion, see “Conversion Metrics” on page 87. • Retain more visitors by segmenting those most likely to return Once you’ve persuaded visitors become customers, you need to retain them as loyal, returning customers. It typically costs 5-10 times more to acquire a new customer than to keep an existing one. WebTrends allows you to evaluate the effectiveness of your loyalty campaigns such as customer newsletters by how recently and how frequently visitors are coming back and engaging in repeat business. Now you can measure whether or not you are increasing the average lifetime value of your visitors. For more information about retention, see “Retention Metrics” on page 99.
What the Business Person Wants to See
Business people need to optimize the effectiveness of their marketing expenditures. They need to run campaigns that drive qualified traffic to their web sites. They make decisions regarding spending more money on tactics that work and reducing the amount on less-efficient areas. The decisions regarding the acquisition of visitors are some of the most important decisions that business people make because the process of acquiring visitors (such as creating ad campaigns, using marketing resources, outsourcing some areas) is expensive. Acquisition data can help you determine if your marketing tactics are successful. With WebTrends you can easily get reports on valuable metrics that reveal how many visitors came to your web site, whether they converted to registering or paying customers and how much value they brought to your organization. Acquisition data can also tell you how comparisons perform over time and which customers have the highest lifetime value.
WebTrends provides performance dashboards that lets you benchmark your acquisition campaigns, as is shown in the following graphic.
The Campaign Performance Dashboard allows campaign managers to view how every level of their campaign performs across their acquisition and revenue KPIs. Campaign managers can easily see if customers are responding to specific offers better than others and can make instant decisions based upon their agreed-upon goals for that particular campaign.
The first page that a visitor sees on your web site is called the entry or landing page. This is the most important page in your web site, because it provides the initial impression for your visitors and influences whether they will continue to look at other pages of your site. Entry pages can tell you whether people more often start at your home page or jump to a specific page on your site—usually from a bookmark or link. Consider also that at the entry point to your site, visitors have not yet begun to navigate the pages of your site. This is an opportune time to guide them in the direction you want them to go. From these pages, you can promote areas of your site that you want them to see by putting noticeable links to those areas. In addition, entry pages usually provide good advertising real estate if you sell ad space on your site or promote your own products or services.
74 WebTrends Marketing Lab 2
Basic Entry Page Usage
The Entry Pages report is the lowest-level report, because you don’t know exactly how visitors got there. That is, your visitors could have used an ad campaign, a search engine, or some other mechanism to get to your page. Yet this page may help you to pinpoint the pages on your site to improve. Based on this report, you can find your leading entry pages and improve them in order to represent your company in the best way possible and direct visitors to other pertinent pages of your web site. The following graphic shows a sample report that identifies the first page view of a visitor at a site.
In this sample report, “Pages” refers to any document, dynamic page, or form. Different types of profiles have different default settings for which file extensions qualify a file as a page. “Visits” refers to the number of visits where the specified page was the entry page. A visit is a series of actions that begins when a visitor views the first page from the server, and ends when the visitor leaves the site or remains idle beyond the idle-time limit. Also, in this sample Entry Pages report, the home page or “Welcome Information” page is the top entry page. However, many visitors entered first through the products and store pages. Perhaps many of these visitors entered because of an ad campaign. If so, this ad campaign may deserve more scrutiny, because the company may have spent quite a bit of money on attracting customers via that campaign. The information in the Entry Pages report can indicate how you might want to optimize the architecture of your web site based on where your visitors are entering. It can also help you determine which external links are most effective. You may want to consider updating META tags and links.
Advanced Entry Page Usage
You can make your entry pages useful by creating specific landing pages for each campaign and making sure that each landing page is not linked to anything except the specified campaign. That is, only the intended campaign should link to the landing page—nothing else on your web site should link to it. Then the landing page redirects the visitor to the page that you want them to view.
By themselves, entry pages are not that interesting, but if you can design them into your web site for analysis, then you can determine who came to your web site because of a particular campaign.
Collecting the Right Data
Web sites can employ a variety of mechanisms to drive traffic to a specific web site and track its success. In general, visitors will find your site through: • • • • Ads (banner, email, traditional media) Searches Links Directly through bookmarks
The following subsections discuss the mechanisms that help visitors find your site. These mechanisms are referrers, ad campaigns, search engines, and email marketing efforts (such as newsletters).
Just as a doctor receives new patients from a referring source—such as another doctor or a current patient—a referrer, or referring URL, is the page on another web site that linked visitors to your site. Referring URLs tell you where your visitors came from to get to your site. You can use this information to determine which external sites are the best ones to place links on, or ads for, your site. This information can also convince you to develop or maintain positive relationships with these sites so that they will continue to offer a link to your site. How do you determine what the referrer is? The record in the data file contains the page that was visited before the page represented by a particular entry. So you can ascertain the referrer for each page from the record in the data file. But more interesting is what initiated a visit to the site. How do you determine the referrer for the visit? This is done by taking the first hit in the visit, looking at that hit’s referrer, and calling that the visit’s referrer. Therefore, all of the referrer’s URLs come from the first hit of the visit. The following graphic shows the domain names of sites that refer visitors to your site.
76 WebTrends Marketing Lab 2
From this sample report, you can get basic information. However, if you have several different ad campaigns on Yahoo, this report doesn’t reveal which one is working best. Consequently, the referrer reports provide general, low-level feedback on your efforts. For more specific information, you will need reports on ad campaigns, search engines and email marketing.
Referring Site, Domain, or URL
Some web sites may have multiple links to your site. If you only want to know what site referred the visitor—not the individual pages on the site that contained a link to your site—you would need to strip out all parts of the URL except the site or domain name. Dong so allows you to discover which sites or domains refer visitors to your web site the most. For example, a site, such as www.referrer.com may have several domains, such as search.referrer.com, your-referrer.com, my_referrer.com, ourreferrer.com, and referrer.com. All of these domains are do referring work for the main site www.referrer.com. Each domain name may have several IP address. For example, 18.104.22.168, 22.214.171.124, 126.96.36.199, 188.8.131.52, 184.108.40.206, 220.127.116.11 might all be used by the www.referrer.com domain. So your reports can give you the site name, the domain name, or the URL information. It depends on the level of information that you want.
The Self-Referrer Issue
Using the referring URL to determine how your visitors came to your site has one major drawback: your own site can appear to be the referring URL. This self-referring circumstance occurs when a visitor begins a visit, leaves your site open in the browser window, stays inactive beyond the 30-minute visitor session window, and then becomes active again. Once the 30-minute threshold is crossed, WebTrends considers this to be a new session; however, this new session will register your site as being the referrer. For this reason, instead of using the referrer page or URL, you may be better off using other means of tracking how visitors come to your site. One useful method is by tracking visits from ad campaigns.
No Referrer - Direct Traffic
“No Referrer” or “Direct Traffic” represents direct traffic to the web site as the result of one of the following: • • • • Visitors who type the domain name directly into the browser Visitors who bookmark the site Visistors who have the page set as their home page Visitors who click on an email link, shortcut, or other direct link
Advertisements can come in many forms, including ads on other sites, popup ads that are triggered, and links embedded in email campaigns. Here are some broad definitions of ads that are frequently used:
These ads include “banner ads” that appear on the web pages of sites that your best prospects are likely to visit. Web-based ads have many forms such as text, moving graphics, a call to action (“Click here to download …”), Flash or streaming banners, pop-ups, and pop-unders.
Acquisition Metrics 77
These ads are directed at publications that your prospects are most likely to be reading. With newsletter ads, you can often choose from among sponsoring the newsletter, sponsoring a column or feature in the newsletter, or placing an ad that will appear among other ads, usually as a text ad.
Campaign IDs and Translation Tables
You can manage your campaigns by using campaign IDs and translation tables to convert the campaign ID into meaningful information. For example, you may see a campaign ID in your data files such as campaign=721. WebTrends allows you to have a text file that (at analysis time) can translate all of your campaign IDs into their corresponding campaign names.
Many ads are designed to initially route the user through a redirect page before they can view the ad content. This redirect page quickly and imperceptibly bounces the visitor to the actual page with the ad content, recording the redirect page as the entry page for the session because it was selected first. (Here, the first hit recognized as an ad campaign in the visitor session is counted.) If each redirect page for each placement is distinct from the others, you can track which version of the ad most often took the visitor to the ad’s content. Let’s say you have two online ads for your product, one on Yahoo, and one on AOL. In addition, you sent an email to potential customers with a link that takes them to the content. If you wish to track them all separately, you would create a separate redirect page for each one. In this scenario, you might have the following pages:
Yahoo Ad: /redirect/yahoo_ad.htm AOL Ad: /redirect/aol_ad.htm Email Ad: /redirect/email_ad.htm
78 WebTrends Marketing Lab 2
By tracking visits to each of these redirect pages in the top entry pages, you can see which ad placements most effectively bring people to your site. The following graphic illustrates the redirect process for WebTrends using the web server data collection method. Remember that client-side tagging will not give you this information unless the redirect page has the proper script. For more information, see “Drawbacks of Client-Side Tagging” on page 23.
Using this illustration, if you looked in the web data activity file you will see a two-step process: The first web data activity file entry:
GET YahooAd.htm - 302 - yahoo.com
This took the visitor from the Yahoo.com to the Yahoo redirect page (YahooAd.htm). Status code 302 means that you were redirected. The second web data activity file entry:
GET PromoAd.htm - 200 - YahooAd.htm
This took the visitor from the Yahoo redirect page (YahooAd.htm) to the promotion ad (PromoAd.htm). Status code 200 means that you were successful.
The following graphic shows a sample report showing the top referring pages.
In this sample report, “Page” refers to any document, dynamic page, or form. Keep in mind that different types of profiles have different default settings for which file extensions qualify a file as a page. Any URL containing a question mark is considered a dynamic page. If “Direct Traffic” is 100% of all your traffic, then your web server is probably not logging the “referrer” field in your data files. You can use WebTrends to create a campaign profile and track either entry or referring pages. However, some ads have several possible referring pages with long, complicated URLs. As a result, it can be more difficult to look up and define a referrer when you set up a campaign profile.
Tracking Multiple Campaigns
Re-direct pages are great for handling a handful of campaigns, but if you’re doing hundreds or thousands of campaigns, this method is impractical. Instead you should use a parameter field containing a parameter ID. The ID can be used to identify all of the attributes that make up the campaign, such as site name (for example, MSN, Yahoo), program (3rd quarter ProductName upgrade), offer (25% off), creative type (120x120 GIF banner), creative (race car image), and so forth. You can then use a translation file (via WebTrends script or custom table lookups) to create reports on which attributes are most effective (for example, did the race car image do better than the Flash movie of a tornado, or was the 25% off offer more effective then the free year of support).
Offline Acquisition Techniques
Entry and redirect pages are handy for off-line acquisition techniques. For instance, if you place an ad in a newspaper or magazine telling people to go to your site, you might get them to type www.YourCompany.com/UpgradeOffer, but you are unlikely to get them to type www.YourCompany.com?CID=C42-61AF.
80 WebTrends Marketing Lab 2
Search engines play a large role in acquiring visitors. Whenever someone uses a search engine, there is the chance that they will use a keyword that triggers links to your web site. Search engines typically come in two flavors:
Paid Search Engine
You pay a fee for every person who clicks through to your site, and you have to monitor which keyword phrases are bringing you the best visitors. With Paid Search Engines, you need to evaluate the effectiveness of money spent.
Organic Search Engine
You pay nothing for visitors who come to your site. You monitor which keyword phrases are bringing you the best visitors. With Organic Search Engines, you evaluate the effectiveness of time spent. People use search engines when they don’t know the name of your site or have no other direct link to click or distinct URL to type in their address box. Web site designers go to great pains to figure out how to get recognized by these search engines and appear in the “top 10” list that appears when a search is performed. Research consistently shows that more than 80% of web visitors use search engines to find what they need. The longer users are online, the more likely they will use search engines and make purchases. Since most web users believe that those sites that show up in the top of the listings are the most important sites, you must take every reasonable measure to make sure your site ranks highly with search engines for the search keywords and phrases that your most valuable prospects use. If you can’t get good rankings by optimizing, you can always try pay-per-click advertising options, which most of the search engines offer. But search engine technology constantly changes. What you did to get a search engine to effectively recognize your site or page today can have marginal results only a few months later. And each search engine has its own proprietary method of creating a result list based on the search phrases or keywords a web user enters. These lists use the search keywords and phrases to create a list of what they interpret as being the most relevant sites. Search engines also use a host of other factors, including how often visitors click on the link to your site from within their list, and how many of the more popular sites containing related content have hyperlinks to your site. Most search engines also let you register with them, and by paying them to place your site in their index, you can get more exposure than if you’d left it up to chance to get noticed. The element that you have direct control over in this mix is making sure that the keywords you planned for visitors to use to get to your site actually make your site appear in the search results. By reviewing the top keywords or search phrases entered by visitors, you can find out if those keywords are driving people to your site. If not, you can modify your web page content to promote your site with search engines—based on those keywords. Some common ways to modify that content involve including the keyword or phrase in the description and keyword meta tags, and increasing the frequency with which you use the word or phrase in the HTML title, a headline, and first few paragraphs of the page. These methods will improve your chances of being found and promoted by a search engine.
Search engine optimization is not the focus of this guide. Please consider WebPosition (www.webposition.com) for a complete discussion of this constantly changing topic. You can use WebTrends to find the search engines that are used most often by visitors to arrive at your site. You might want to register with search engines if you find that your site is not being noticed.
With WebTrends you can generate reports on organic search engines (non-paid search engines) and paid search engines. The following graphic shows a sample report about most recent search engines.
82 WebTrends Marketing Lab 2
WebTrends allows you to compare this information with information from a report on the most popular phrases for your site. The following graphic shows a Most Recent Search Phrases (All) report.
Using the information from this report and the previous report you can compare search engine rankings with the popularity and competitiveness of phrases to get a complete picture of how the web site is performing. Search engine rankings allow you to understand where your site shows up in the list of search results for certain phrases; for example, if you have a phrase that performs particularly well in terms of conversion, but your search engine ranking is low, you may want to try for more highly qualified traffic by boosting your ranking. WebTrends can also analyze paid and organic search engine usage and generate reports that show the total effectiveness of your search engine marketing and optimization strategies based on activity, depth and duration of visit. You can receive separate reports on paid search engine, or organic search engine, or both.
When you want to reach prospects directly, consider using direct email and either your own customer database or a purchase marketing list. You can also email to your in-house list of registered visitors, who have opted-in to receive communications. By using email marketing, the recipient can click on a link to your web site, and this visit is automatically recorded and catalogued by WebTrends.
You can use WebTrends to track email campaign results using entry/landing pages as a primary or complimentary metric to the other measures produced by email solutions. WebTrends can help you to determine how far recipients get into the conversion process, as well as what they do once they’ve completed the process and on subsequent visits. Advanced email solutions will track clickthroughs to the site, campaign conversions and revenue—and in some cases visitors’ clickstreams/paths—but this is where the overlap with web analytics solutions ends. Unless the visitor’s activities are tied directly to the campaign, meaning the visitor entered your site through the link contained in your email, viewed campaign details/pages, and converted on the campaign offer, most email solutions will not measure it. You can make your entry pages useful by creating specific landing pages for each email marketing campaign and make sure that each landing page is not linked to anything except the specified campaign. That is, only the intended email marketing campaign should link to the page—nothing else on your web site should link to it. Then the landing page redirects the visitor to the page that you want them to view. To analyze the detailed interactions your email visitors have with your site beyond summary campaign information such as the number of responses and conversions, you will need a WebTrends solution. If visitors left campaign-centric pages, where did they go? What content groups or products (beyond the one featured) most interested them? Did email recipients purchase products that weren’t featured in the campaign? All of these questions can be answered by using WebTrends. The following graphic shows a report that provides information about all types of campaigns, including email marketing.
84 WebTrends Marketing Lab 2
This report lets you compare different kinds of campaign types to see which are the most effective. Of course, the effectiveness is related to how much money you are spending on each campaign.
Tracking Multiple Email Campaigns
You can use re-direct pages for handling a handful of email campaigns, but if you’re doing hundreds or thousands of campaigns, this method is impractical. Instead you should use a parameter field containing a parameter ID. The ID can be used to identify all of the attributes that make up the campaign, such as site name (for example, MSN, Yahoo), program (3rd quarter ProductName upgrade), offer (25% off), creative type (120x120 GIF banner), creative (race car image), and so forth. You can then use a translation file (via WebTrends script or custom table lookups) to create reports on which attributes are most effective (for example, did the race car image do better than the Flash movie of a tornado, or was the 25% off offer more effective then the free year of support).
Acquisition is the most expensive step in getting visitors to your web site. Monetary expenditures on advertising, search engines, newsletters, and similar campaign efforts often make up the largest share of a company’s budget. But without visitors—especially qualified visitors—your web site is meaningless. Once you have customers, you can work on converting and retaining them. Fortunately, conversion and retention are far less expensive.
Finding the Features in WebTrends Products
The following paths will help you to find the topics discussed in this chapter in WebTrends.
Entry Pages and Referrers
Click Web Analysis > Report Configuration > Campaigns > New Campaign
Click Web Analysis > Report Configuration > Campaigns To create a report about ad campaigns, Edit a sample profile and click Visitor History. Make sure that Campaign History is checked.
Click Web Analysis > Report Configuration > Custom Reports > Reports or Dimensions To create a report about search engines, Edit a sample profile and click Visitor History. Be sure to check Search Engine History.
Acquisition Metrics Worksheet
Use the following worksheet to help understand how you want to acquire visitors. Consideration
Will you be using ad campaigns to drive traffic to your site? How many of these campaigns do you need to track? Do you complete test campaigns before you begin the real ones? Note: Test campaigns can help you understand which campaigns work the best. Will you use redirect pages? Do you intend to outsource the creation of your ads and the serving of your ads? Will you track referrers? Are you relying on statistics from organic search engines? Are you using paid search engines? Will you use a email newsletter campaign?
86 WebTrends Marketing Lab 2
After you have attracted visitors to your web site, you can measure how often the visitors take an action in line with what you intended. In other words, conversion means getting visitors to do what you want. For commercial web sites, conversion usually means how often visitors convert into paying customers. However, many commercial sites are interested in “lead generation” in which a sales lead may generate a potential conversion to a paying customer later. In either case, the metrics involved with conversion measure the process by which you persuade visitors to take the actions that you intended for them to take. Your conversion rate is a measure of your ability to persuade your visitors to take those actions. The following scenarios are examples of conversion: • • • • • • Visitors purchasing products Prospects registering for more information Customers using your self-service section Investors downloading your annual report Employees using your internal site to schedule vacations Visitors registering for the site’s newsletter or to enter contests
The conversion process may involve several steps through your site as visitors navigate their way. Conversion analysis helps you evaluate which types of content successfully support conversion.
First-time Visitors vs. Repeat Visitors
Conversion is not the process of doing, rather it is the process of a non-doer becoming a doer. Consequently, you may want to filter across visitor segments to see what first-time visitors and first-time buyers do rather than what repeat visitors do. This means running a filter on a profile and doing some custom table filtering. Getting a new visitor to convert is a sign of success. The following graphic shows a report comparing the number of visits by new and returning visitors to your site.
Conversion is the beginning of the rewards for having spent so much time and money on the acquisition step. Retention (discussed in Chapter 9, “Retention Metrics” on page 99) involves the process of how you minimize the ongoing cost. It is much cheaper to keep a customer happy than to get a new one.
Understanding Navigation Measurement
Navigation measurement is one of the most fascinating areas of web analytics. You can theorize about why certain things happen on your site, but to draw any firm conclusions, you need to understand how visitors use your site by the paths they take within it. Knowing how visitors navigate your site can help you determine what types of content interests your visitor. It can also help you identify trouble spots that may have caused visitors to exit your site. Understanding where visitors go on your site helps you answer questions such as: • • Did your visitors only view the top-level pages, or did they delve a little deeper to see details about a certain topic? Where on your site are people running into dead ends or backtracking? You know that the fewer clicks a visitor has to make to get to the information they want, the higher their satisfaction with their overall experience. Consequently, you want to make sure that they aren’t having difficulties locating information. When people go to the Contact or Support sections, what seems to be driving them there? Are they tending to come from certain areas of the site that you should examine more closely? Where and why are shoppers deviating from the “ideal” straight-through checkout process that you created? Should you change the order of the steps or provide certain information earlier in the process? Is your site design causing visitors to go in circles? If people abandon your site before getting the information or doing the transaction that you designed for the site, why are they leaving? Observing their first few clicks into the site—or into each section—can help identify the pages that need to be examined for confusion, inconvenience, lack of information, poor visual appeal, or other obstacles. Can you get similar information from looking at the last few pages of aborted visits? You have an idea of what constitutes a typical or ideal visit, but are you oversimplifying? Are there really several kinds of visits? What are they? Is your site designed to work well for many kinds of visitor “missions?” Are you ignoring the needs of an important group of visitors? You would like people to visit more of your site than they do. Where are the best places to encourage visitors to explore new parts of the site?
Path Analysis answers the question “Where do visitors go on my site?” Path analysis helps you determine whether visitors navigate your site the way you expected them to, and if not, where they go instead. Path analysis can also help you track movement between pages, or can leverage your content group settings to track movement between defined content areas. Different approaches to path analysis provide different types of insight into your visitors’ activity. You can take a free-form approach and track the top paths starting with the entry page. This analysis lets you know where visitors began and where they went on your web site. Or you can look at the most popular routes on your site. You can also narrow or focus your approach by examining certain hot spots on your site, examining which paths led visitors to hot spots and which paths followed from the hot spot. WebTrends excels at path analysis, providing comprehensive information about the navigation of visitors on your web pages.
88 WebTrends Marketing Lab 2
Analyzing Complete Paths
A complete path means that you track all the pages that a visitor traverses during a visit session. This is virtually the same as manually examining each hit collected and analyzed. If you took this approach, you would have so much data to interpret that you would never be able to recognize patterns in that data. Also, the amount of data your web analytics solution would have to process would tax your server’s performance considerably. So how can you narrow down the data on all of the paths?
Analyzing Focused Paths
Typically, you know the pages that are significant to your organization. Rather than tracking all visitor paths through your site, just track the paths to and/or from significant pages such as entry pages, exit pages, the home page, search pages, and key conversion pages. Doing so would narrow down the scope of how much data you analyze, providing far more focus than you would get by tracking every page. By narrowing your focus, you can research deeper. Consequently, you can track to the path depth that you want. On anything other than a simple site, you may still encounter so many paths to or from a given page that meaningful patterns in visitor behavior may be difficult to discern. It is also possible that certain paths, although technically different, have the same content. Sometimes it is not always intuitive to look at the progression of pages along a path and easily understand exactly what that behavior indicates. For example, instead of seeing visits to the Wireless phones View page, you may want to see the level of interest in visits to all product detail pages. In this case, you can use Content Groups to group related product detail pages.
Analyzing Complete Content Group Paths
By grouping together pages that are equivalent indicators of visitor behavior, you can track broader patterns as visitors traverse a complete path through the various content groups you have defined. In other words, you are applying meaning to a group of hot spots and the directions that visitors take in getting to or leaving the hot spot. But much like tracking the complete path through pages, interpreting your results can be confusing due to the volume of results. Once again, to obtain information that is far easier to handle and interpret, it may be best to focus on specific content group paths.
Analyzing Focused Content Group Paths
A focused content group path is the select list of content groups, in order, that a visitor traverses in arriving at, or departing from, a particular content group. The results you get from this type of tracking offer extremely high levels of insight into how visitors use your site. Content groups allow you to ignore visits to pages that are of no interest by simply omitting the page from any content group. If you are interested in seeing whether visitors move from the Store Product Page to Accessories to Ordering, or from the Main Catalog Page to Specific Product Information and then to Warranty information just before Ordering, you can ignore side trips to the Glossary page or Investor Relations page. The ultimate value of the content group method depends on the skill with which the content groups and their member pages are chosen. Part of your success depends on selecting the right groups and the right pages for each group. The groups must be comprehensive enough to simplify the picture, but not be so comprehensive that they contain within themselves patterns that should be exposed.
Focused content group path analysis is an excellent way to classify visits, which can be the basis for a sophisticated redesign. Because most or all of a visit can be captured in a good content group path analysis, it is possible to see whether the different functional parts of your site defined by the content groups tend to be viewed together. For example, if the Technical Information section of a site is visited far more often by people who visit a particular product section, and not by other visitors, it may make sense to add better links between these two sections or to beef up the technical content of the product information.
Tracking the Path Most Traveled
You probably want to understand where a visitor is most likely to go after viewing a specific page or content group. You may also want to know what page or content group most often preceded a visit to a specific page or content group. This is called single jump analysis. This type of analysis shows you if your visitors are going where you expect them to go. If they are not, you would want to look for obstacles that might be preventing them from following the path you want them to follow. By ensuring that people visit specific areas of a site, you can be sure that these areas have the opportunity to succeed.
A more specialized case of path analysis is scenario analysis. This type of analysis helps you discover if people are visiting all the pages in a scenario that you intended for them to visit. You typically have an interest in seeing them complete the steps in the scenario because completion of the scenario often translates into revenue. By telling WebTrends the pages that make up a scenario, you can track how many people started the process and where along the way they dropped out. If dropout rates are significantly higher on specific pages, you may consider factors such as poor site design or insufficient information on those pages. Scenario analysis also allows you to exclude from analysis any irrelevant pages that the visitor visits while completing the scenario. This is something that would not be possible if you were simply tracking a specified path through the site. The following is an example of one of the most commonly used web site scenarios—an online purchasing scenario, commonly called a shopping cart. The typical shopping cart scenario might include the following steps: 1. Open the shopping cart. 2. Add products to the shopping cart. 3. Start the checkout process. 4. Complete the order. The scenario analysis technique tells you what percentage of visitors who complete one step in the sequence also complete the next step. An obvious example is shopping cart completion, but the technique can be applied to a variety of other scenarios, including applications for services, storefinders, feedback forms, personalization processes, and some kinds of on-site searches. The following graphic shows an Purchase Conversion Funnel report with entry and exit pages.This view shows from where people entered the scenario, and where they went to when they exited the scenario at particular step, or where they abandoned the scenario. For instance, when a visitor leaves a step, visits another page (page X), then leaves the site, page X is shown as the exit page from the last scenario step. Note that in this report: • On the left-hand side, you will find the entry pages that lead to one step in the funnel. For more information about entry pages, see “Entry/Landing Page” on page 74.
90 WebTrends Marketing Lab 2
On the right-hand side, you will find the exit pages that show where you visitors went when they left that step in the funnel. For more information about exit pages see “Exit Page and Exit Ratio Analysis” on page 92.
In this example, the largest number of customers dropped out of the process after opening the shopping cart. Only a little more than 28% of customers who started a shopping cart actually added an item to the cart. Interpreting these results depends on many variables. Whether or not a visitor starts a process, such as a purchase, is often more dependent on merchandising issues and perceived value than on site design. In contrast, whether or not a visitor finishes a process once they have started it usually depends on variables such as clarity or convenience. These variables are well within the control of the site designer. For this reason, scenario analysis of individual processes is an excellent tool for evaluating the effects of changes in the design of a process. After you configure WebTrends, analysis can be done on a before and after basis. Note that in the table that accompanies the funnel graph, the “Scenario Analysis Step” column lists the names of the steps in the defined scenario. Each step marks progress on the path that is being monitored. The Step Conversion Rate is the percentage of visits converted from the previous step in the scenario. Scenario Conversion Rate indicates the percentage of visits converted from the first step in the scenario. Often the nature of scenarios is non-linear, meaning visitors may enter a step out of sequence. For instance, with a “Quick Checkout” process, a visitor may be able to jump from step 1 directly to step 4, and would never be counted in steps 2 or 3. Also, in the case of a visitor leaving the site at step 2, then returning later in new visitor session to that same step, this may cause the number of step 2 visitors being greater than those of step 1. WebTrends allows you to view these “Step Transitions.” This view focuses on how visitors proceeded from one step to the next or through the scenario. If a visitor proceeded directly from Step 1 to Step 3, Step 3 will appear among the pages listed to the right of Step 1. You should be careful in selecting the pages for your scenarios, so that you can determine problems. It pays to think through possible problem areas and to try using those pages as steps in the scenario you want to analyze. For example, you might find that visitors are abandoning your site at the page in which they are asked to state their address. Or they might be dropping out at the page that requests their financial information.
Another part of the conversion process takes place after visitors have found their way to a page containing an internal search feature. Visitors can use this search mechanism to find items on your site. Consider stores such as Powell’s, Amazon, or Barnes & Noble that have an internal search for books (and other items). By examining the keywords and phrases that visitors were searching for, you will learn what your visitors’ interests are. This information reveals explicit, rather than inferred, implied interest. You now know the words that your visitors are using to describe your content. This information can help you better organize your site, and it can help you to optimize your use of external search engines.
Exit Page and Exit Ratio Analysis
So now you understand various ways that people arrive at your site and some of the conclusions you can draw—based on how they got there. But what can you learn by knowing the exit page, the last page visited in a visit session? Leaving your site can be viewed as a failure of site design if the top exit pages were not where you expected your visitors to exit. Determining the positive versus the negative value of leaving via a specific page is relatively subjective, but it can suggest what on your site works, and what doesn’t.
92 WebTrends Marketing Lab 2
The following report shows a sample report of last pages that visitors viewed before leaving a site.
The visit-to-exit ratio compares the number of exits from a given page to the number of visits to that same page. It is important to know what percentage of visitors to a page leave directly from that page, because pages that receive the most exits are almost always the most visited pages. To create this ratio for all of your site’s pages, simply start with the most important areas on your site. After you have calculated the ratios, you can review the pages with the highest percentage of exits per page view to prioritize the exit pages. This kind of information can often reveal a key page with a high visit-to-exit ratio that does not appear among the top exit pages.
A dead-end path is a path in which the visitor goes from one page, to another, then returns to that original page. Dead-end paths can be both good and bad. In some cases, it can mean that visitors were looking for specific information, assumed that a given link would take them to that information, but upon arrival at the new page, realized that they had not found what they were looking for. This activity means that they are having trouble finding information.
A dead-end visit can just as easily mean that the visitor followed a path out to its natural conclusion, and then came back to the previous page to continue looking for other information. A simple example of a good dead-end path can be seen with an online news site. The person opens the main page, clicks on the International News section, and then clicks on a specific article. After reading the article, they return to the International News section to select another story. This is exactly how you would expect these pages to be used.
Gleaning Demographic Information Through Registration Forms
Many sites require users to fill out a registration form when they reach a point in which they need to download some content or access more in-depth information on the site. These sites typically request varying levels of personal information too, depending on how much their audience is willing to reveal. Often, there is a delicate balance between collecting valuable information and alienating your visitor. Some web sites request information regarding gender, age, income, and a zip code. This allows the visitor to remain anonymous, yet still provides the web site owner with valuable demographic information. However, many sites do request more detailed information about the visitor. It just depends on what the site owner is trying to achieve by collecting visitor data. But how does visitor information get tied to an individual hit if there’s no authenticated user field to tie together hits by the same visitor? And where does the visitor information entered in the forms go? Just as you did in sessionization, you can identify the visitor by using a cookie ID, the authuser field, or the IP address. Now let’s explore where the visitor information goes. Most online registration forms use the GET method of requesting content. With this method, information entered in the form can be attached as query parameters in the data activity file. There are two ways that these query parameters can then be used to capture visitor information, and they depend on the type of system you have set up to process your web activity data files—a web analysis program or a web data warehouse.
The GET method has a limit of 2000 characters. The POST method can also be used, but the content can’t be seen in the data activity files. Therefore, the GET method is preferred. In one method, WebTrends parses the hit (in the web activity data file) for the visitor information parameters you specified that it should locate. WebTrends then takes that information and enters it into a database. With each new hit, the software checks the visitor identifier against visitors already in the database. If the visitor identifier is new, it adds a new row and adds visitor information to that row. If the visitor already exists in the database, the program attaches the hit information to that visitor record.
94 WebTrends Marketing Lab 2
The other method involves the use of a web data warehouse, such as the WebTrends Marketing Warehouse, a database that is designed to hold visitor information. You tell the warehouse which parameters hold specific web visitor information, and the warehouse processes the web data activity file, captures the visitor information, and stores it in a visitor database table within the warehouse. All behavioral information associated with that hit is also tied to the visitor using the visitor ID. Subsequent hits go through the same process. If the ID in the hit matches a visitor that has already been identified, only the behavior information for that visitor is updated. If the visitor has not yet been identified, then a row is added to the visitor table, and all the behavioral information from that hit is associated with that visitor.
For more information about warehouses and the WebTrends Marketing Warehouse, refer to “Integrating and Exploring Data” on page 107. Keep in mind that any issues you would encounter using cookie IDs or IP addresses to identify the visitor in visit sessionization, will also occur when using those same items to identify visitors.
Evaluating Visitor Behavior by Browsing Your Site
WebTrends SmartView displays a page from your web site fully rendered—as it appears to visitor—and annotates this page with results and metrics from analysis. In a companion window, SmartView displays the page’s metrics with reports for Page, Paths, Scenarios, or the Entire Site. This display makes it easy to evaluate the popularity of each individual page link with click-through, path, and scenario metrics superimposed on the page you are viewing. You can use SmartView to analyze page performance, providing insight into page conversion, path analysis, and overall web page statistics such as unique visitor counts.
The following graphic shows a typical SmartView page of the Zedesco web site.
With SmartView you can get a sense of where your visitors are going and relate the traffic to the actual visual appearance of the page. Consequently, you can see relationships quickly—even ones you did not anticipate. This may lead you to rethink the page’s design or direct you toward new territory for further analysis. You might also want to use SmartView to double-check a hunch or an assumption. Since SmartView presents a higher-level and immediate view of the data, you probably will not use SmartView to publish reports on a weekly basis.
Once you’ve told WebTrends how to identify visitors so that you can associate visitors with their behavior on your site, you can track the paths that those visitors take through your site. In fact, you can track the distinct pages they traverse through your site, and you can use your content group settings to track how they navigate through your site in terms of the types of content they viewed. Tracking pages can be useful in some cases, but typically you are more interested in getting a bigger picture of how visitors use your site. For this reason, you may prefer tracking paths through content groups rather than through pages.
96 WebTrends Marketing Lab 2
Finding the Features in WebTrends Products
You will find the topics discussed in this chapter in WebTrends.
Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Path Analysis or Web Analysis > Report Configuration > Path Analysis
Click Web Analysis > Reports & Profiles > Edit a profile > Advanced > Scenario Analysis or Web Analysis > Report Configuration > Scenario Analysis
Web Analysis > Report Configuration > Scenario Analysis To create a report using shopping carts, Edit a sample profile and click Visitor History. Make sure that Purchase History is checked.
Click Web Analysis > Report Configuration > Custom Reports > Dimensions To create a report about search engines, Edit a sample profile and click Visitor History. Make sure that Search Engine History is checked.
Use the following worksheet to understand how well visitors are converted on your site. Consideration
Identify the top 5 key pages in your site that you want to see traffic moving to. What are the paths moving to and from those pages? Identify the scenarios (especially any registration or checkout pages) in your site. If you have an internal search feature, do the most popular keywords and phrases really fit your product? Are there other words that visitors should use? Should keywords be listed on a search page or other pages to help visitors make the associations you want them to make? Identify your dead-end pages. What is the meaning of each dead-end page? What kind of program can you set up to periodically measure the conversion rate to see if improvement has occurred?
98 WebTrends Marketing Lab 2
The vast majority of web sites need to retain their visitors. You’ve gone through a lot of hard work and expense to attract visitors and convert them into buyers or registered users. Now it’s time to keep those visitors. From a monetary perspective, retention involves the process of how you minimize the ongoing cost. It is much cheaper to keep a customer happy than to get a new one. Customers who return again and again have the highest value, which translates into profits for commercial businesses. To make retention work for you, you must find out more about your visitors and their behavior. Understanding your visitors and their behavior can answer the following questions: • • • • • • • On which visitors should you spend marketing dollars? When? What can you expect in future sales from your existing visitors? How do you predict which ads and products generate the best visitors? What kind of incentives should you provide to get a visitor to do something you want them to? Can you predict which visitors will be responsive to your program? Should some visitors be contacted more often than others? How can you put a value on your visitors and business as a whole, and project this value into the future?
Visitor retention activities are an investment—with the expectation that the value of the investment will rise. But initially you’ve got to know more about your visitors and their behavior.
Visitor Segmentation and Behavior Segmentation
By grouping, or segmenting visitors along lines such as gender, age, income, or location, and then comparing web activity between these population segments, you can learn a lot about whether you’re reaching your intended audience. This is where visitor information gets correlated with behavioral information in visitor segmentation. That is, the who (visitors) becomes correlated with the what (their behavior). Behavior reflects what the visitors did. Which content groups and directories did they look at? What kinds of searches did they do? Who your visitors are and information related to them (demographics, referrers, entry point, browser, time of visit) is called visitor space. What your visitors do is called behavior space. Any slice of information relating to visitor is called visitor segmentation. Any slice of information relating to visitor behavior is called behavior segmentation.
The following graphic shows the relationship of visitor space and behavior space.
Once you have identified the behavior of specific population segments on your web site, what then? This level of insight into your web visitor allows you to take action, if needed, to better capture the audience you want to attract. This is the information that lets you implement a continuous improvement cycle—you measure the activity for a given offer or ad campaign, make a decision based on that measure, take some action based on the decision, then you re-measure to see what effect the action had. Let’s consider what might happen with a scenario in which a wireless phone company uses a cellular phone package to target 18 to 25-year-olds. The company might run an advertisement that web visitors access through promotions on ten different sites. These ten web sites were chosen because they are sites geared toward a younger crowd. When visitors link to the ad, before learning more about the package, they are prompted to fill out a survey that requests information on their age, sex, zip code (if applicable), and current occupation. After one week, the cellular phone company reviews which referring sites tended to send the greatest number of 18 to 25-year-olds–the target audience. At that point, the company continues paying for the promotion on sites that referred the most targeted visitors, but discontinues the ad on those sites that failed to do so. By tying web behavior to their web visitors, the cell phone company was able to quickly identify where their marketing dollars were effectively being spent, and where they were wasting their money. Even if you only learn about the behavior of visitors, you can move ahead. For example, you can compare the repeat rate of visitors generated by different banner ads or keyword phrases.
Number of days since the most recent visit of a visitor. Note that zero recency means that the visitor returned within less than 24 hours. Most businesses find recent customers to be more valuable than customers whose activity has been dormant for a long time.
Number of visits since the visitor was first tracked. There’s a great deal of difference in value between a 100-time repeat visitor and a 2-time visitor.
Number of days between visits for visitors. Note that zero latency means that the visitor visited every day. Latency can be especially helpful for businesses where orders and contacts have a defined cycle (for example, a subscription-based business and businesses selling durable goods or high ticket items). All three measurements can be used to determine the potential value of your visitors.
100 WebTrends Marketing Lab 2
Lifetime value is a concept that applies to commercial web sites, because these sites need a long-term gauge for their repeat customers. Lifetime value represents the total sales generated since tracking a specific visitor began. The following graphic shows the lifetime value of visitors to the Zedesco web site.
Reports that reveal lifetime value have a great influence on the types of offers you might present your visitors. For example, the report in the following graphic shows the lifetime value of buyers for the most recent campaign they responded to, and displays it in a drilldown. A drilldown enables you to examine this information at a highly summarized level, and navigate to successively more detailed levels of campaign data; for example, viewing lifetime value of buyers by demand channels, partners, marketing programs, marketing activities, campaign IDs, campaign descriptions and more.
If you run this report again a few months later and find that the average latency for most of your customers is increasing, then you should take action to correct this behavior.
WebTrends allows you to collect the behavior of individual visitors over a period of time. This feature is called Visitor History, and it is primarily used to track the activity of visitors’ purchasing behavior such as how well visitors have responded to advertisements, how much money they spent, how many times they bought something, and how many items they bought.
General Information about Visitor History
WebTrends stores a record of information per visitor. So, for every visitor, there’s a set of information recorded each time the visitor views a page. Each time the visitor returns to that page, WebTrends can compare the current activity with past activity and measure various attributes for that visitor such as:
Lifetime count of purchases from shopping cart
Most recent purchase value
The value of the most recent purchase
Days before first purchase
The number of days between a visitor’s first visit and first purchase
Days since first purchase
The number of days since a visitor’s first purchase
Days since most recent purchase
The number of days since a visitor has purchased an item In other words, Visitor History allows you to measure visitor activity according to recency, frequency, latency, and lifetime value. Visitor History can help you to find out which customers you might lose. For example, the information you get from Visitor History might prompt your marketing departments to send special offers to customers who haven’t been active for a while. In general, Visitor History can help you to convert one-time users into frequent users. The Visitor History records are stored in the Visitor History database, which is “under the hood” of WebTrends where you don’t see it or have to worry about it. The only thing you have to do is make sure that you activate the Visitor History check box in the UI if you need Visitor History for some analysis. For more information, see “Finding the Features in WebTrends Products” on page 105.
Specific Information about Visitor History
Visitor History stores a set of attributes for each visitor. Each time a visitor visits your site, WebTrends analyzes the attributes, comparing new information with older information. The following is a complete list of attributes that are stored in the Visitor History database: • • • • • • • • • • • Number of hits Number of visits Time of first visit Time of last visit Total number of seconds of visit time - added up from all lists of that visitor Entry URL from a visitor’s first visit Referring URL from visitor’s first visit Referring URL from visitor’s first visit in which he/she bought something Most recent referrer for a buying visit The first ad campaign that brought the visitor to the web site The most recent ad campaign that brought the visitor to the web site
102 WebTrends Marketing Lab 2
• • • • • •
The total of all the money that visitor spent on your web site over a lifetime The total number of purchases made by a visitor The time that the visitor made his/her first purchase The time that the visitor made his/her most recent purchase The search engines used by the visitor to get to your site The search words/phrases used by the visitor to get to your site
WebTrends stores aggregated information about purchases. This aggregation is sophisticated enough to make fine distinctions such as invoice rejection. For example, if a visitor goes to a shopping cart site and accidentally submits twice on a purchase page, WebTrends can detect the unintended action and make sure that it is counted once instead of twice. WebTrends can also detect an accidental bookmark to a purchase page and count that visit properly.
Example of Visitor History Usage
There are many ways to use Visitor History to help retain your customers. Here are some examples. • Products and visitors with highest lifetime value Compare which products are being viewed by visitors with the highest lifetime value. To retain your most valuable visitors, you could send them special offers that are associated with the products they are most likely to purchase again. • Recency and lifetime value Compare recency with lifetime value and determine if some of your most recent buyers are ones with the highest lifetime values. If over a period of time you see that some of your most valuable customers are dropping off in their purchases, then you might make them a special offer. • Amount of time between first visit and first purchase Run a report to find the time of the first visit of some customers and then compare that with the time of their first purchases. Your goal is to reduce the amount of time between that first visit and the first purchase. • Referring URL (or ad campaign) and lifetime value Run a report to list your top referring URLs (or ad campaigns) in relation to lifetime values of visitors they bring to your site. You might consider identifying the top three referring URLs (or ad campaigns) and work with the organizations that own them to increase your referrer rate. • Demographics and lifetime value Compare demographics and lifetime value to see what kinds of people have the greatest lifetime value. Such factors as age, gender, income level, and geographic location may indicate whether you should increase marketing efforts to one group or another. • At-risk visitors/customers To find out about past visitors who have not been to a site in a number of days, you can use the recency metric and then decide if you would like to appeal to them (perhaps based on previous loyalty) with special offers.
WebTrends Marketing Warehouse lets you take visitor history and segmentation to a higher level. With WebTrends Segment you can define segments based on a characteristic of a web site event, visit, or visitor that is captured in the Marketing Warehouse and use this information to fuel targeted marketing campaigns. With WebTrends Segment you can: • Define segments on the fly Create new segments while working with your data and immediately apply the segment to your exploration. • Identify opportunities Once defined, segment performance can be dynamically measured with an easy-to-use data exploration tool that seamlessly queries and drills into any marketing initiative or business event on the fly, so you get a complete picture of segment performance. • Target visitors Once you’ve identified a segment of visitors ideal for a repeat marketing campaign, you can quickly drill into the details and pull a list of subscribers, export the list to your email system and start measuring results. Customer segmentation starts with the business event or performance data you want to analyze; whether that is a conversion scenario, a campaign event or a specific customer segment, you have complete control to break down the analysis in real-time by any of the business events and visitor profile data being collected in the Marketing Warehouse. With custom segmentation, you can answer the following types of questions: • • • Which customers are most likely to buy within the next week? Which customers are highly engaged and what content are they looking at? What products do my highly valuable customers look at?
The following graphic shows an example of the powerful segmentation and exploration capabilities that are available in WebTrends Explore:
104 WebTrends Marketing Lab 2
Unique Visitors, Unique Buyers
The purpose of your web site is to present information to people and, usually, to encourage them to take some action such as purchasing. Hits and visits provide measures of what and when those people are viewing, but your real target is the people behind those actions. If you know how many of each type of individual come to your site, you can develop a strategy for changing the visitor’s behavior or for changing what you might offer them. It is at this point that identifying and counting unique visitors comes into play. In order to track unique visitors, you first need a means of unambiguously identifying each visitor. As discussed in Chapter 4, “Visitor Identification” on page 31, cookies and authenticated user names are the best solutions to this problem. Although unique visitors and unique buyers refer to the individual visitors to your web site, keep in mind that one unique visitor may view any number of pages on your site within the framework of a visitor session. Therefore, 1,000 unique visitors can generate 50,000 page views.
Tracking Unique Visitors with WebTrends Analytics
WebTrends Analytics counts uniqueness by keeping track of daily unique visitors, weekly unique visitors, etc. by using a cookie. The following graphic of the Visit Summary from the Visitors Dashboard shows a tabulation of unique visitors over a 24-hour period.
After you define your unique visitors, you may be interested in certain groups of these visitors, such as those who have a lifetime value of at least $500.
Tracking Unique Visitors with WebTrends Marketing Warehouse
With the WebTrends Marketing Warehouse, you can track unique visitors over any period of time. Supported by a powerful relationship database, the Marketing Warehouse allows you to get to the who behind the visit, and it lets you define business events that help you to understand and communicate with your target audience. Using WebTrends Segment you can look at unique visitors who have a recency of once a day or once a week, and compare their lifetime values, or you can create your own custom segments. By combining business events with these customer segments, you can create targeted customer lists based on the preferences and behavior of your customers and drive relevant marketing campaigns to these customers.
Finding the Features in WebTrends Products
Retention metrics are enabled using Visitor History. To access Visitor History settings, edit a profile and click Visitor History.
Use the following worksheet to understand how well your site retains visitors. Consideration
On which visitors should you spend marketing dollars? When? How often? When launching ads, do you target specific visitors or send out general information to all visitors? Which visitors will be responsive to your programs?
Which visitors should be contacted more often than others?
How can you put a value on your visitors and business as a whole, and project this value into the future?
106 WebTrends Marketing Lab 2
Integrating and Exploring Data
So far, this guide has discussed what is often called data farming. That is, you figure out what you want to examine, and then you configure WebTrends to review those specific areas of interest. Just like a crop, you harvest these same pieces of information over and over again on a schedule. Doing so enables you to understand trends in visitor activity in response to changes you’ve made to your site. But what do you do if you have existing customer data that you would like to correlate to their web behavior? Or how do you investigate a feeling that one dimension relates to another, or that several dimensions correlate significantly with each other, and you want to find out whether your intuition is correct? At this point, you need the help of a web data warehouse and a tool that allows you to report from the web data warehouse. A web data warehouse integrates the data that you want to explore. A warehouse also lets you connect external data to your web behavior, and export your web behavior to external data. External data, for example, could be information from a Customer Relationship Management (CRM) system or a customer database (with customer demographics). Using a web data warehouse gives you flexibility in analysis and reporting. WebTrends lets you look at your data in a number of different dimensions simultaneously. To view reports from the web data warehouse, you can use Microsoft Excel or another reporting tool. With Excel, you can make use of its PivotTable function to view and compare data in two dimensions (2D). You can also make graphs based on two measures as the X and Y axis. For more information about the Excel reporting solution, see “Deeper Reporting and Exploration” on page 110. In general, a web data warehouse and an associated reporting tool (such as Excel) require more people, resources, and knowledge-power.
Data Integration and a Web Data Warehouse
It’s important to understand the difference between using a web data warehouse with WebTrends for analysis and reporting rather than relying only on WebTrends. A web data warehouse contains a database specifically designed to store web activity and web visitor data. Unlike the summary tables used by reporting tools in WebTrends, a web data warehouse stores detailed data rather than accumulating and summarizing it into daily, weekly, monthly, quarterly or yearly tables, and then discarding the raw hit data. A web data warehouse uses a series of tables to capture and store web activity data. A warehouse has a hit table with IDs that can be tied to other tables containing hit data from processed web data activity files. Hit data is analyzed to create a visit table and some other tables with visit-specific information available in the hit such as the referrer for the visit, an ad campaign, or a content group. Each visit table record has a visit ID along with several IDs that allow it to match a given visit to the appropriate records in those related tables. The visit tables associate the web activity in a hit with a specific visit session.
A web data warehouse also provides tables that hold visitor information—first name, last name, gender, age, email address, phone number, zip code, customer number—any information you ask your web visitors to provide about themselves that they’re willing to enter. These tables contain visitor IDs that are associated with visit information. Now you can perform queries on the database to correlate specific visitor attributes. Perhaps you might correlate age or gender with a particular web behavior, such as a visit to a particular ad. Consider the ad for the cellular phone package example, previously discussed in “Visitor Segmentation and Behavior Segmentation” on page 99. You could examine visits to the ad that originated from a given referring site made by visitors aged 18 to 25.
Tying Data to External Databases
You can further enrich the data you have about your visitors by tying the data in your web data warehouse to external data sources such as demographic data. The key is that your web-related data and the external data sources must have some variable in common so that you can match records from your web data to your external data. Because the web data warehouse is in a database form, it is straightforward to join to an external database. Some warehouses have a mechanism by which you can join your web analysis results to an external source and then present that data in a custom report.
Perhaps you have the state associated with each web visitor record, and you want to tie that activity into a database that describes demographics by state. Numerous databases exist that can help you segment your visitor population. For example, WebTrends GeoTrends provides demographic information. Let’s consider a straightforward scenario: Zedesco’s budget limits them to airing a TV commercial in only one state. If they are using their web site as a basis for deciding in which state to air the commercial, what information might they need? One of the most basic pieces of data they could look at is which states show the most web viewing activity, such as the most page views or the most visits. If two states show similar activity levels, the next step might be to see which state has the most buying power. To do this, they could tie into a demographic database that contains information on average income level by state. If they find that between the two states showing the most activity one has a lower average annual income, then assuming all other variables are equal, they’d air the advertisement in the wealthier state.
Joining web visitor information to web visitor activity is useful for marketing professionals as they try to more accurately target their marketing using the web. But you can also use your web activity and web visitor data for account management. You do this by joining the web activity of individual web visitors with their account contact data in Customer Relationship Management (CRM) systems such as Siebel Call Center or PeopleSoft. CRM systems are database-driven applications that are generally used to manage the information about an organization’s prospects and customers. These systems often contain information about customers or customer prospects, such as: • • • • Correspondence Contact information Previous transaction information Communication via email, phone, or regular mail
108 WebTrends Marketing Lab 2
Joining web visitor and web activity data to complex databases such as those used by CRM systems requires the structure of a web data warehouse. To join the two sets of data, you need one or more shared keys, or IDs, to match the records in one database with records in the other. Typically, this will be some visitor ID in the web activity database, and a customer ID in the call center database. Other possible shared keys between the two databases could be combinations of first and last names or email addresses. The following graphic illustrates the shared keys between two databases.
Joining web activity with visitor information lets salespeople understand their visitors’ interests with information such as: • • • • • Which web pages they visited How many times they visited those pages How long they stayed Which products or topics they researched How much information and interest they have about specific products as evidenced by the white papers, demos, or other marketing and technical materials they downloaded from the web site
Service professionals can also use this combination of information to review a customer’s web activity to prepare them for handling the customer’s issue. Useful information includes troubleshooting topics, frequently asked questions, or technical white papers that the customer has already examined. In addition, by reviewing how often specific troubleshooting topics or frequently asked questions are accessed, support organizations can determine if products or documentation have weaknesses or other issues that need to be addressed.
Integrating and Exploring Data
The following graphic shows an environment that is running machines that use web analysis and warehouse data. In this illustration, the client machine is able to view reports on the warehouse using a reporting application such as Crystal Reports. The warehouse can communicate with other sources of data, such as CRM or Enterprise Resource Planning (ERP) and wed that information with the warehouse data.
Reporting from a Web Data Warehouse
While a web data warehouse provides an effective vehicle for organizing and storing your web data, it often doesn’t provide a means of reporting on that data. To view your web activity data from a warehouse, you need to use a reporting tool, such as Microsoft Excel. Here are the steps to use Excel to report from a web data warehouse. 1. Export the WebTrends data to Excel. 2. Export data out of the web data warehouse to CSV format 3. Import CSV-formatted data to Excel. It is important to note that the imported data be in the CSV file format. Also, Excel has a limit of 65,000 rows of data.
Deeper Reporting and Exploration
WebTrends allows you to move beyond standard reporting to dig deeper into your analysis and compare several different variables with each other. If you find that your reports do not fully cover what you want to examine, or do not show the data from the perspective you wish to view it, you can create new reports “on the fly.” You can do this kind of exploration by exporting WebTrends reports to Excel spreadsheets, called SmartReports, and then working with Excel’s PivotTables. Through SmartReports, you can also develop graphs and charts that correspond to the tables of data (using trend data).
110 WebTrends Marketing Lab 2
SmartReports for Excel is more than an export of the WebTrends data into Excel. In addition, SmartReports creates an Access database containing the WebTrends data and includes WebTrends data integration features from within Excel. Some of these features allow for automated updates of newer WebTrends data, joining external data to WebTrends and other features. Excel reports themselves include additional features, such as a graphical interface that lets you join external data to WebTrends. You might use for SmartReports to verify whether a correlation between several variables exists so that you can then structure your web analysis to generate periodic reports on those variables and track them over time. You can also use SmartReports to combine web analytics data with external data, such as marketing cost or product cost, to calculate GMROI. After you have used Excel to reveal specific gross margin trends, you can track your variables over time and chart them in SmartReports for further insight. For example, you can calculate gross margin trends and chart the sum of gross margin revenue by campaign for insight into which campaigns are most successful for you.
With Excel, you can drill down in the report to discover more critical pieces of information. This capability can be especially useful when you are dealing with a hierarchy within the dimensions you’re analyzing. For example, to analyze campaign data, each category, such as Demand Channel, might have a subcategory, and within that subcategory, you might have a further division. The following table shows how this might look: Category
Subcategory Level 1
Subcategory Level 2
Marketing Program Biz Rate 2003
Subcategory Level 3
Marketing Activity Datafeed Electronics Store
My Simon 2003
Datafeed Electronics Store
Within WebTrends reports, you can interactively click on a given dimension and drill down to the next level. For example, if instead of examining all product categories (Camping, Hiking, and Boating) you only wanted to view information about the Hiking category, you could simply click on the Hiking Product category, and view information about Boots, Clothing, and Backpacks. Within SmartReports, you can drill as far as you have specified in WebTrends drilldowns. For instance— using the example above—within the Hiking product category, you could drill down three levels, and examine visits to pages in the Internal Frame subcategory of the Backpacks subcategory.
Integrating and Exploring Data
The following graphic shows a SmartReport with categories and subcategories.
In SmartReports, you can use Excel’s tools to choose the exact dimensions and measures you want to compare, and you can discover significant correlations between dimensions. These tools use automated machine learning and statistics to uncover trends, which Excel can present in a variety of graphs, tables, and charts. Data exploration is an iterative process. You need someone who is adept at statistics and is willing to look at the same data again and again in order to find the “nuggets” in the data.
112 WebTrends Marketing Lab 2
The following graphic shows an Excel chart with trend data mapping campaigns by sum of gross revenue for December 2003. This is an example of charting data that is calculated in Excel and shown in a graphical format.
The following graphic presents another Excel chart of trend data mapping. Note that you can use PivotTable reports to filter the data by group, department, etc., and that this filtering can change the visual representation in the graph.
Integrating and Exploring Data
Another data exploration exercise might involve examining relationships between visitor attribute dataincome level, zip code, gender and the content groups and ad campaigns visited. To do this, you would have Excel compare each visitor attribute and combination of visitor attributes against content groups, against the combination of content groups and ad campaigns, and then against ad campaigns. But practically speaking, what are the benefits of data exploration? Data exploration can be used to reveal significant trends in customer behavior. For example, with an online travel site, women from zip code 97215 with an annual income of $70K visit the last minute deals pages and respond to email ad campaigns more than any other visitor population segment. Knowing this, you might choose to send out a targeted email for a last minute deal, and then use standard web analysis reporting to see if that email campaign is effective.
Overhead and Monetary Costs
Data exploration is much more resource intensive than looking at web analysis data in the standard way. Getting the most results from data exploration requires personnel who can look at all of the possible information that they can mine from your data and understand which correlated segments are worth pursuing. They must thoroughly understand data statistics and data interpretation to make the most of your investment. Another major cost regarding data exploration involves computing power. Data exploration can exhaust computing power very quickly, because you have to do multiple cross tabulations of various dimensions to find which ones correlate. Your web site does not have to register a million hits to justify the cost of data exploration. If you attach a value to your traffic, you should still consider the possible returns on the resources you. Data exploration will give you a lot more insight at a higher (and deeper) level, but the exploration involved can be expensive. You may be exploring many avenues before you reach the right one (for example, using A/B testing); so you’ll need some intelligence to figure out which way to go. Since data exploration is very open-ended, you need to narrow down the many possibilities and achieve meaningful results. Consequently, a data exploration solution for your company doesn’t mean that you merely purchase more software, plug it in and watch your income grow. You have to look hard at adding the right kind of personnel who will work hard to interpret the data.
Using Reports for Continuous Improvement
The purpose of reporting on your web site activity is to have easily interpreted information that allows you to make improvements to your site, marketing campaigns, or other aspects of your business that are tied to your web site traffic. Just as in any continuous improvement cycle, you need to determine your objectives for your site, plan how to implement those objectives, execute that plan, then generate reports that allow you to assess the success of that plan. As you discover what works and what doesn’t, you make small, incremental changes. To complete the cycle, you measure the impact of those changes with other comparative reports.
114 WebTrends Marketing Lab 2
Data Integration and Exploration Worksheet
Use the following worksheet to help understand more about data integration and exploration. Consideration
Do you have external data that you want connected to web behavior? Can you afford a web data warehouse in terms of costs relating to people, software, hardware, and planning? Will there be compatibility issues if you bring any previous-existing external data into the warehouse? Do you have data that you need to investigate in SmartReports? Do you have Excel experts who know how to work with PivotTables?
Integrating and Exploring Data
116 WebTrends Marketing Lab 2
For a scenario or multi-step process, the percentage of initiated scenarios that were not completed during the visit. Scenarios can be defined many ways—for example, the entire shopping process, a finite checkout process at an e-commerce site, a registration process at a lead generation site, or a search process at an information site.
A page that is displayed after a visitor completes an action or transaction: for example, a Thank-you or Receipt Page. An Acknowledgement Page is often important in Scenario Analysis, where it is an indicator of a completed scenario.
The process of attracting a visitor to your web site.
A general term referring to nearly any site measure, including visits, hits, visitors, and viewing time.
A link, usually commercial in nature, consisting of a graphic or text that takes a visitor to a web site when clicked on. An abbreviation for “advertisement.”
A specific effort to attract visitors to your site through ads. It may be one individual ad or a coordinated set of ads treated as one entity for reporting purposes. On the web, ad campaigns usually consist of emails, graphics on other sites or on a wireless interactive appliance, and traditional media such as direct mail, print, broadcast, outdoor advertising, etc. In WebTrends, ad campaigns are set up by the reporting administrator with a unique URL/landing page, a starting date, an ending date, and a cost. Same as Campaign and Marketing Campaign.
A click on an ad resulting in a jump to the site being advertised.
A display of an ad on a page that is viewed during a visit. There may be more than one ad view on a page.
An Internet term loosely referring to the location of a web site or web page on the Internet or the Web. Or, more specifically, an identifier for a specific computer that is connected to the Internet.
Combining data of two or more dimensions in a report. For example, adding up all Departments to get Total Division data. While such combinations are normally sums, any type of formula might be used.
A visitor who used a username-password login process to get access to all or part of a web site. The username (but not the password) is captured in a specific field in web site log files or through client-side data collection tags. Although authenticated username is one of the most accurate ways to count unique visitors, the cookie method is the most preferred.
Technique that limits access to Internet or intranet resources to visitors who identify themselves by entering a user name and password.
A statistical term referring to the sum of a measure divided by the number of items measured. For example, for a series of 11 visits consisting of 3, 7, 7, 7, 8, 10, 15, 22, 25, 25, and 35 page views each, the average number of page views is 14.9 (total 164 divided by 11), the median is 10 (the 6th in the series of 11) and the mode is 7. In statistics, average is also called the mean.
The average of the frequencies of all the visitors during the reporting period, where each visitor’s frequency is the number of times they have visited the site since WebTrends visitor tracking began.
The average of the latencies of all the visitors during the reporting period, where each visitor’s latency is the average elapsed time, in days, between all their visits since WebTrends visitor tracking began.
Average Lifetime Value
The average of the lifetime values of all the visitors during the reporting period, where each visitor’s lifetime value is the total monetary value of a visitor’s past orders since WebTrends visitor tracking began.
The average of the recency values of all the visitors during the reporting period, where each visitor’s recency is the averaged elapsed time, in days, since their last visit.
Measure, in kilobytes of data transferred, of the traffic on a site.
Banner, Banner Ad
An online advertisement, usually a graphic, which can be anywhere on a web page but typically refers to a horizontally elongated graphic of significant size located at the top or bottom of a web page.
In a browser, a shortcut to a web site page that is created by the visitor to allow a quick one-click return to the page in the future. Bookmarks are called “Favorites” in some browsers. Visitors arriving at a site by clicking on a bookmark will appear as a “Direct Traffic” entry in Referrers reports.
A program - such as Microsoft Internet Explorer and Netscape - used to locate and view web pages as well as to follow hyperlinks. The Browser is identified in the “Agent” or “User Agent” field of a web site log or through standard client-side data collection tags.
A specific advertising effort to attract visitors to your site. A campaign may be one individual ad or a coordinated set of ads treated as one entity for reporting purposes. For online channels, campaigns usually consist of emails, graphics on another site or on a wireless interactive appliance, and traditional media such as direct mail, print, broadcast, outdoor advertising, etc. In WebTrends, campaigns are tracked using WebTrends query parameters. Same as Ad Campaign and Marketing Campaign.
118 WebTrends Marketing Lab 2
A “creative” describes the characteristics of a marketing activity, such as color, size and messaging; for example, a “Buy Now” graphic. These creative elements are used to encourage clickthroughs to the web site. Campaign Creative is a level within the drilldown categorization scheme set up by the WebTrends administrator, which allows for reporting on groups of campaigns in a way that is meaningful to the report users.
In certain WebTrends reports, a drill-down feature allows the user to navigate from a highly summarized level of data to successively more detailed levels of data, organized along a concept hierarchy. With Campaign Drilldown, users can examine visits, page views, revenue, average order size, and more, by Campaign Partner, Demand Channel, Marketing Program, Marketing Activity, Campaign Name, Campaign Creative, Campaign Offer, and other campaign attributes.
A unique campaign identifier used to calculate campaign success, which may involve several different marketing activities, or a single effort. Campaign ID is a level within the drilldown categorization scheme set up by the WebTrends administrator, which allows for reporting on groups of campaigns in a way that is meaningful to the report users.
This is a user-defined category, which might include online banner ads, e-marketing newsletters, and direct mail campaigns. Campaign Type is a level within the drilldown categorization scheme set up by the WebTrends administrator, which allows for reporting on groups of campaigns in a way that is meaningful to the report users.
The page or series of pages viewed when a visitor goes through the process of buying something online.
WebTrends can use Child Profiles to report on a web site that shares a log file with other unrelated sites due to a constraint or choice by a hosting provider. Child profiles can be helpful if an ISP or web hosting service hosts multiple customer sites on their web servers. To a web site visitor, a customer’s site can appear as a distinct, stand-alone domain, but often the web activity data for each customer site is recorded and lumped together in the service provider’s main web server log file. If service providers want to offer their customers a set of basic web activity reports with data specific to each customer’s site, they need a means of breaking out data by customer. Because service providers also want to reduce management and maintenance of this data splitting process, they want WebTrends to auto-discover and split out these data subsets while parsing the log file. Parent-Child profiles provide this auto-discovery functionality, and also creates profiles, called Child profiles, for these data subsets.
The act of activating a hyperlink, usually by physically pressing down (clicking) on a mouse button when the cursor is over a link on a page. In Web advertising, a click is an instance of a user activating an advertising link to go to an advertiser’s web site or page.
The number of clicks on an ad as a percentage of the total views of the ad during the reporting period.
A computer (or software on a computer) that accesses resources provided by another computer, called a server.
An error occurring due to an invalid request by the visitor’s browser. Client errors are in the 400 range (see Status Code on page 134 for a list).
Client-side Data Collection
An alternative to traditional web server log file analysis that involves collecting data directly from the visitor’s browser (the client) rather than from server log files, improving data accuracy. Special script in a page’s source code is used to transmit page-level data, not “hit-level” data, to a data collection server, dramatically reducing data volume and decreasing processing time. Client-side data collection obtains more accurate information than log files do—by accurately tracking visitor activity normally hidden by browser’s local cache and proxy and caching servers like those used with an AOL account—as well as by collecting extra, customized data not included in normal web server log files. Accuracy is also improved since spiders do not trigger client-side tags; with log files, spiders can appear to be “real” visitors unless their activity is filtered out. However, client-side methods provide no information on server technical performance or bandwidth use. WebTrends’ proprietary client-side data collection technology is called SmartSource.
Combined Log File Format
A basic (“common”) log file with two additional fields, the Referrer and User Agent fields. Also referred to as Extended Log File Format.
A group of one or more web pages that is treated as one entity in certain reports such as Content Groups and Content Paths. Content Groups are tracked using WebTrends query parameters.
A consecutive sequence of two or more Content Groups viewed during a visit.
Content Management System (CMS)
Software application that enables organizational information to be developed collaboratively. Web sites are commonly created and published using a CMS.
Action taken by a visitor that is key to measuring web site performance. Can also be a goal that the visitor wants to achive using the site.
The percent of a group (of visits or visitors) that took a specific action of interest. Conversion can encompass the entire visit population, such as the percent of all visits that involved a completed registration. Conversion can also refer to a very small and precise action, such as the percent of people at step 3 of a scenario who continued to step 4; or it can apply to a subpopulation, such as the percent of knowledgebase searches that result in issue resolution.
120 WebTrends Marketing Lab 2
A hit or visit filter created in the Custom Reports feature of WebTrends Administration. Custom filters can be a variation of a filter already in use or can be completely new, based on a variety of hit or visit characteristics. Visit-related custom filters (such as Entry Page, Referrer, and Campaigns) are especially powerful, allowing the inclusion or exclusion of entire visits as defined in the filter.
A customizable WebTrends report consisting of summary information—usually graphs—from individual WebTrends reports in a profile, all grouped on one page. Dashboards provide a quick overview of key information for individuals, departments and specific roles.
A destination page is an administrator-specified page used in Destination Paths reports as the page to which all the analyzed paths lead.
Elements or categories being reported on in a WebTrends report. A dimension usually does not have a numerical value; for example Pages and Content Groups. They are statistically described using Measures—which do have a numeric value—such as visits, views, view time, etc. In WebTrends reports, the dimension is the first column or the first two columns if both a Primary and Secondary dimension are used. Dimensions are also presented in drill-down format in some WebTrends reports.
A web site is made of files that are usually separated into groups of similar files, such as all product pages, or all Human Resources pages. In a complex web site, directories can contain smaller directories, such as Human Resources procedures pages and Human Resources job listings, and the levels of directories can go quite deep. Directories are often reflected in the address of a web page, which includes not only the name of the page (joblistings.html), but also the series of directories it belongs in separated by slashes (/international-company-info/USA-company-info/USA-humanresources/). WebTrends uses directories two ways. First, it is possible to filter page views by specifying directories to include or exclude. Second, a Directories report tallies the activity in individual directories.
A line item in the Referrers reports that pertains to visits that have no known referring site, domain, or URL. Usually, this means that visitors arrived at your site by typing the URL of your site into their browser address window, they used a bookmark, or they clicked on a link in an e-mail. If “Direct Traffic” is the only line in a Referrers report, this usually means the Referrer field is not used in your traffic logging.
DNS Lookup (Domain Name Service Lookup)
The process of converting a numeric Internet Protocol address into a text domain name. For example, DNS Lookup will convert the IP address 255.255.255.255 to the domain name YourDomain.com. DNS Lookup can be turned on and off by the WebTrends administrator. Also, GeoTrends relies on DNS resolution to provide state, country, and company information. “DNS” refers to Domain Name Server. DNS Lookup is also called IP Resolution and Domain Name Lookup.
A legacy term referring to pages that were defined as “documents” by the system administrator. Traditionally, a page is a document if the content is static, such as an HTML page.
The text name corresponding to the Internet Protocol address of a computer on the Internet. For example, webtrends.com is a domain name. A domain can be associated with many IP addresses but an IP address can have only one domain.
A broad categorization of domain names identified by the suffix, such as .edu (for domains related to educational institutions), .com (for domains related to commercial web sites), .org (for domains related to non-profit organizations), .gov (for domains related to governments), and many others. The domain type does not necessarily reflect the true nature of the web site, as domain suffixes are only loosely regulated, if at all.
In certain WebTrends reports, the drill-down feature allows the report user to navigate from a highly summarized level of dimension-related data to more detailed levels of data, organized according to the hierarchy of a dimension. On a web site, “drilling down” is the act of going further down a branch of the site in search of more detailed information. Often, drilling down results in seeing a series of different navigation bars, each appropriate to its own level.
A page that is created by the web server from a template, or a general page structure, which is filled in with content pulled from a database. Servers “build” dynamic pages from particular components according to requests they receive from browsers. The URLs of dynamic pages typically consist of the template name, followed by a question mark, followed by the content for the displayed page as a series of text strings separated by ampersands in the format parameter=parametervalue. For example, a page showing a blue Empire couch might be /product.asp?item=couch &type=Empire&color=blue. The parameters can be of great interest in web analytics, when shown as tabulated summaries of views of couches, Empire items, and blue items, or combinations of these.
The first page, file, or content group in a visit.
The first file requested in a visit. A visit has one and only one entry file. Files may be of any type, including a page file.
The first page requested in a visit. A visit has one and only one entry page. Note that a visit will have no pages if it doesn’t include a page file.
A page view that is both the entry and the exit page; the only page in a Single-Page Visit.
The last page viewed in a visit.
A collection of information stored under a unique name, often in the form “name.extension” where the extension identifies the type of file and, usually implies what kind of program can open or view it. On the Web, common types of files are: page files (.htm, .asp, .jsp, .cfm, and so on), image files (.gif, .jpg, .png, and so on), applet files (.js, among others), non-page document files (.doc, .txt, .pdf, and so on.), and style files (.css, among others). While a page file is technically different from a page (see Page on page 127), a page will always include a page file.
Corresponds to a file’s extension. For example, a file named graphic.gif is identified as type gif.
122 WebTrends Marketing Lab 2
A setting in WebTrends that instructs the program to exclude or include (to the exclusion of all else) certain visits or hits from the analysis. In WebTrends, filters can be used individually or in groups, and individual filters can be combinations of different subparts.
A cookie that is created by the web site you are currently visiting.
A visitor who has made his or her first purchase. Also called New Buyer.
Scripted pages that pass variables back to the server. These pages are used to submit information entered by visitors in the form’s fields.
The number of times a visitor has visited a site since tracking with persistent cookies and Visitor History began. Average Frequency is the average of the frequencies of all the visitors during the reporting period. Frequency is a retention metric and is part of RFM (recency, frequency, monetary) analysis. If visitors did not visit the site during the report time period, their frequency is not included.
File Transfer Protocol. A standard method of sending files from one computer to another over the Internet.
Attrition that happens as site visitors go through a scenario, a series of defined steps such as a purchase or a registration on a web site. Because the number of people participating in each step is usually smaller than the step before, a graph of the declining participation resembles a funnel.
In certain WebTrends reports, a drill-down feature allows the user to navigate from a highly summarized level of data to successively more detailed levels of data, organized along a concept hierarchy. With geography drilldown, users can examine activity by areas of visitor origination, for example, viewing visits, page views, revenue, or average order size, or viewing by Region, Country, State/Province, or City.
The optional GeoTrends Database resolves Internet Protocol addresses of visitors into more meaningful data such as the region, country, state/province, city, area code, designated marketing area, metropolitan statistical area, and time zone data corresponding to the location of the owner of a specific domain name. In the specific case of AOL IPs, location is resolved to geographic regions served by AOL as opposed to the location of AOL in the state of Virginia. GeoTrends Database replaces the older WebTrends’ Company Database.
A graphics file format and file extension (*.gif) commonly used on web pages, referring to Graphics Interchange Format.
A value in a WebTrends report indicating that Internet security software such as Zone Alarm Pro, Symantec’s Norton Internet Security, or a proxy server modified a web request to mask the referring URL.
A request for a file by a browser. Since “file” refers to images, styles, and many other elements besides HTML pages, a single web site page view can involve dozens of hits. Because the number of hits is so heavily influenced by the complexity of a page, hits are a far less helpful measure of site traffic than visits or visitors. The hits statistic is somewhat useful in assessing the load experienced by a web server. WebTrends SmartSource Tags do not capture hit-level data.
The main or introductory page of a web site, usually designed with the expectation that it is the first page a visitor sees. It is also the default page that is sent in response to a request containing only the domain name.
Home Page URL
The URL for the home page of the site analyzed in the report. The home page URL is specified during WebTrends setup in order to help WebTrends consolidate hits to several versions of the home page, for example, flash- and non-flash-versions or framed and frameless versions.
The abbreviation for Hypertext Markup Language, which is used to format text files so that web browsers can display text with appropriate hyperlinks, font sizes, and other text formatting.
The abbreviation for Hypertext Transfer Protocol, a standard method of transferring data between a web server and a web browser. HTTP is the text string that appears at the beginning of web addresses, and it informs a browser that the request is for a web page as opposed to an FTP site or another type of browser destination.
See “Session Termination Time Frame” on page 132.
Instrumented Web Page
A web page that contains a WebTrends SmartSource Tag. The SmartSource Tag does two things. First, it transmits traffic data (similar to that in a standard IIS or Solaris log) to the WebTrends SmartSource Data Collector for processing into reports. Second, if set up to do so, it also collects and transmits a wide variety of optional extra data to the same Data Collector.
A numeric phrase used to identify a computer connected to the Internet. IP addresses consist of four one-to-three-digit numbers separated by periods, for example, 18.104.22.168. WebTrends allows filtering activity coming from a specific IP address or range of addresses.
An abbreviation for Joint Photographic Expert Group, referring to a compressed graphics format common on the Internet. Also called JPG.
Navigation or moving from one page to another using a link.
124 WebTrends Marketing Lab 2
A page on a web site—which may or may not be the home page—where the visitor arrives. For example, in an email campaign, you would use a landing page as the page to which the email directs the prospect via a link.
The average number of days between visits for a given visitor since tracking with persistent cookies and Visitor History began; for example, those who visit on average every 7 days. For a given visitor, a lapse of 12 days between the first and second visit, and a lapse of 24 days between the second and third visit, equals a latency of 18 days. Note that a zero latency means the average time between visits is less than 24 hours. If visitors did not visit the site during the report time period, their latency is not included.
Increases the capacity of a web sites hosted on multiple servers, called a server farm. Load balancers distribute page requests to the appropriate server, to even distribute the traffic. They can also monitor services running on each server, and distribute requests to the server that are functioning properly.
The total monetary value of a visitor’s past orders since tracking with persistent cookies and Visitor History began. Average Lifetime Value is the average of all the Lifetime Values of the visitors who visit the site during a reporting period. If visitors did not visit the site during the report time period, their Lifetime Value is not included.
On a web page, text or an image that has been coded to take a browser from one page to another, or from one site to another.
A file on a web server that contains records of activity related to requests for site content from browsers, spiders, and other outside entities.
Log File URL
The full address, including network ID, drive and directories, of the web server log files that are to be analyzed in a profile.
A visitor who visits a site relatively frequently.
Same as Lifetime Value; see page 125.
A specific effort to attract visitors to your site. It may be one individual ad or a coordinated set of ads treated as one entity for reporting purposes. In the web world, marketing campaigns usually consist of e-mails, graphics on another site or on a wireless interactive appliance, and traditional media such as direct mail, print, broadcast, outdoor advertising, etc. In WebTrends, campaigns are set up by the reporting administrator with a unique URL/landing page, a starting date, an ending date, and a cost. Same as Campaign and Ad Campaign.
A statistical term referring to sum of a measure divided by the number of items measured. Also called the average. For example, for a series of 11 visits consisting of 3, 7, 7, 7, 8, 10, 15, 22, 25, 25, and 35 page views each, the mean number of page views is 14.9 (total 164 divided by 11), the median is 10 (the 6th in the series of 11) and the mode is 7.
Quantities being reported on in a WebTrends report. Measures are quantitative in nature and appear in WebTrends reports as columns to the right of the Dimension column(s), statistically describing them. In Custom Reports, the WebTrends administrator can define and use a wide variety of Measures.
A statistic used as an alternative to Average. In a collection of numbers that have been ordered by size, the Median is the middle value. It is smaller than exactly half of the numbers and larger than the other half of the numbers. The Median is less distorted by extreme numbers than is the Average. For example, for a series of 11 visits consisting of 3, 7, 7, 7, 8, 10, 15, 22, 25, 25, and 35 page views each, the median is 10 in this series (the 6th in the series of 11). The average is 14.9 and the mode is 7. For an even numbered series, such as 12 visits, the median is the average of the middle two numbers.
A statistic used as an alternative to Average. In a collection of numbers, it is the number that appears most often. For example, for a series of 11 visits consisting of 3, 7, 7, 7, 8, 10, 15, 22, 25, 25, and 35 page view each, the mode is 7. The median is 10 in this series (the 6th in the series of 11), and the average is 14.9.
The total value of a visitor’s past orders or transactions since tracking with persistent cookies and Visitor History began. Same as Lifetime Value. Average Monetary Value is the average of all the Lifetime Values of the visitors during a reporting period. If visitors did not visit the site during the report time period, their Monetary Value is not included.
Most Recent Campaign
The last campaign that a visitor responded to since tracking with persistent cookies and Visitor History began. For the report time period selected, all conversions and other activity are tracked and attributed to visitors’ most recent campaigns. Only those most recent campaigns whose durations have not expired are included, and the report administrator sets this expiration. Thus, even if the conversion does not happen on the first visit generated by the most recent campaign, the appropriate source is “credited” with the conversion. If visitors do not visit the site during the report time period, their most recent campaign is not included.
The domain name or Internet Protocol address of one of the sites in multi-homed log file. You can report on a single domain using the Multi-Homed Domain Filter.
Multi-Homed Log File
A single log file that contains the access information for multiple web sites. To specify which domains are analyzed in this type of file, use the Multi-Homed Domain Filter.
Multi-Homed Web Server
A single server that hosts more than one web site.
A visit in which more than one page was viewed. In other words, any visit that is not a single-page visit.
The act of moving from location to location within a web site, or between web sites, accomplished by clicking on links. Navigation also can refer to the overall structure of the links on the site, comprising the paths available to the visitor.
A visitor who has never been to the site since tracking with WebTrends and persistent cookies began.
126 WebTrends Marketing Lab 2
New visitors are identifiable only on sites that give out persistent cookies. WebTrends identifies visitors as new visitors if they have no site cookie when they arrive, and they are able to accept a cookie for their subsequent page views. If they already have a site cookie when they arrive, they must have been to the site before. In a log file, a new visitor’s first page view has no cookie, but all other page views do. It’s important to realize that “never been to the site before” can be evaluated only for the time period during which the persistent cookie has been given out. In fact, when a persistent cookie is first implemented, all visitors appear to be first-time visitors.
See “Direct Traffic” on page 121.
A purchase consisting of one or more items.
The number of completed purchases.
The number of items purchased in an individual order.
The monetary amount of an order.
Organic Search Phrase
A search phrase for which your site shows up on result pages, because of the search engine’s method of ranking pages as opposed to paid placement.
This is a term appearing at the bottom of WebTrends report tables for any table that spans several pages. In these situations, “other” refers to table line items that appear on the other pages of the table, whether before or after the portion of the table being viewed. WebTrends uses the “other” quantity to indicate the proportion of the total picture that is the viewable part of the list.
Paid Search Phrase
A search phrase for which your site shows up on result pages due to paid placement with the search engine as opposed to its method of ranking pages (Organic).
Same as “web page.” In terms of a web site visitor’s experience, a page is a unit of site content, often resembling a paper page of indefinite length and width, that has a single URL address. What the visitor sees as a “page” is usually a collection of files, always including one page file (.htm, .jsp, .asp, .cfm, and so on), plus, depending on the page, image files (.gif, .jpg, .png.), style files (.css, among others), applet files (.js, among others), and a variety of other types of files. In WebTrends default settings, a page is defined as a file with an extension such as .htm, .asp, .jsp, or .cfm. This definition can be modified by the administrator to include or exclude any file extension.
Technically, a page that is displayed by a browser. This term is often used loosely to also include page files that are delivered to a browser, whether or not they are displayed on the screen. An example of a Page View that is not actually displayed is a Redirect Page.
A program used on a Palm device to display site content, similar to Netscape or Internet Explorer on PCs.
A portable personal computer small enough to fit in the palm of a person’s hand, specifically those made by Palm and using the Palm operating system.
Parameters are located in the URL immediately after a question mark and are followed by an equal sign and a return value, known as name=value pairs. For example, in the URL /products/ furniture.asp?cart_id=445& product=couch, there are two name-value pairs. In the first, cart_id is the name and 445 is the value. In the second, product is the name and couch is the value. When URLs contain more than one parameter value, name-value pairs are separated by an ampersand (&).
A specialized way of setting up profiles for different web sites that share servers and log files. Setting up a Parent-Child arrangement automates the creation of profiles and reports on a number of domains or subdomains from a single log file. New domains or subdomains automatically generate new profiles.
The sequence of all pages viewed during a visit, or any portion of that sequence. In WebTrends reports, paths either have a designated starting point (the visit entry page or a designated path start page) or a designated end point (“destination page”); or, paths are Top Paths, which, regardless of specific start page or end point, are common routes through the site. Technically, any visit contains many paths, each consisting of two or more sequential page views. Paths can also refer to content group paths instead of paths consisting of individual pages. The length of paths tracked is either determined by the number of pages viewed, or by the path analysis length limit if the number of pages viewed is greater than the limit.
A report displaying and quantifying paths that fit the criteria set up by the WebTrends administrator including a starting point or an ending point (destination), and a path analysis length limit.
Path of Interest
Describes a concept and practice of focusing path analyses on a particular area of interest. With WebTrends this is typically done with Destination Paths and Paths From Starting Page reports, though technically Top Paths and Paths From Entry are also paths of interest.
In a comparative date range display, a positive or negative percentage that indicates the size of the increase or decrease between the first and second date range. A value of 100% indicates that the second date range’s value is twice that of the first date range’s value; that is, 100% more than the first value. Percent change is calculated by subtracting the first date range’s value from the second date range’s value and dividing the result by the value of the first.
A cookie that lasts longer than the duration of a visit and is saved in the Cookie folder of a browser’s computer. It is used by WebTrends to distinguish new from returning visitors among other things.
The operating system, such as Linux or Windows, used by the visitor’s computer.
A specific good or service that is sold or displayed on a web site.
This is the highest-level categorization of products used in product drilldowns, for example Electronics. The WebTrends administrator defines levels used in the categorization scheme to allow reporting on groups of products in a way that is meaningful to the report users.
128 WebTrends Marketing Lab 2
This is a collection of WebTrends report settings and definitions used to generate, analyze and distribute the set of reports. It is integral to producing WebTrends reports. The characteristics of a Profile include the location of the log files and specific information about their content that will be used in analysis, such as which page URLs are to be assigned to Content Groups and which page URLs are to be starting pages for path analysis. When specified in conjunction with a Template, the Profile determines a complete report configuration that can be analyzed. A Profile can have several templates, just as a template can be applied to many Profiles. A web site can have one or many Profiles and templates.
An established method of exchanging data over the Internet.
Used to build customer segments based on attitudes, values, beliefs and opinions as opposed to the “factual” characteristics of demographics. Political views, learning patterns or music tastes would qualify for psychographic segmentation. Marketing research usually combines demographic and psychographic information to build a more comprehensive understanding of customers. Because the Internet is still a relatively new and evolving medium, one which the mass market is still getting used to and whose usage patterns are determined both by levels of Web experience and type of person, psychographics are of great interest for the Web. The ability of an online broker to convert browsers to online traders, for example, will depend to a large degree on the type of person using the site: are they confident people who like to ‘give things a go’ or are they risk-averse followers of the masses? Psychographic segments built on attitudinal and behavioral characteristics will often be good indicators of how customers will use and react to a web site.
A completed transaction involving an exchange of money for a product, service, privilege, or other item.
Purchase Conversion Funnel
A specific kind of scenario analysis consisting of steps leading to online purchases. The steps of the scenario are designated by the WebTrends administrator.
An individual piece of a query string consisting of a parameter name and a value for the parameter.
The part of a URL that contains information about the content of a dynamically generated page. Web servers use this information to retrieve the specified content from a database and combine it with a template to display a page. A Query String can also contain information that is not directly used to construct a page, but which is intended for use in reporting or other functions. WebTrends’ SmartSource SDC tagging is often used to insert valuable reporting information into the query string. In many dynamic URLs, the Query String is the part of the URL that follows a question mark.
The number of days since a visitor’s most recent visit since tracking with persistent cookies and Visitor History began. Zero recency refers to a visit in the preceding 24 hours. Average Recency is the average of the recency of all visitors during the reporting period. If visitors did not visit the site during the report time period, their Recency is not included.
A web page that is coded to take the visitor’s browser to another page automatically and usually immediately. Many redirects are instantaneous and the visitor does not see the redirect page. Some have time delays and allow the visitor to see the redirect page for a certain number of seconds. Redirects are used to help track clicks that go off site, or to an executable, downloadable, or other file that cannot normally be logged.
A web domain, site, or page that contains a link to one of your site pages that was used by a visitor to get to your site.
A web domain that contains a link to one of your site pages, used by a visitor to get to your site. For example, yahoo.com.
The URL of a specific page on a site that contains a link to one of your site pages that was used by a visitor to get to your site.
Registration Conversion Funnel
A specific kind of scenario analysis comprised of steps leading to online registration. The word “funnel” refers to the typical attrition of visitors from one step to the next. The steps of the scenario are designated by the WebTrends administrator.
Visitors who bought something during the reporting period and are known to have bought something previously as well. Use persistent cookies to track Repeat Buyers. If buyers have cookie parameters for purchases from your site dating from their purchases during the reporting period, they are repeat buyers. Visitors whose browsers do not accept cookies appear as “unknown” in reports that display first-time vs. repeat buyers.
A term loosely applied to graphs and a table associated with an individual analysis, or the collection of all such reports resulting from the analysis of a given profile and template.
Report Period, Reporting Period
The dates covered by the data displayed in a report. WebTrends users may select a report period of any day, week, month, quarter, or year, or a custom date range and can switch between date ranges as desired.
A set of report characteristics consisting of content, the content’s order of appearance, graphic type specification, style, format, language, and other settings which determine the form and content of a finished report. A given profile can have many templates assigned to it, and the report user can view different templates depending on permissions in place. Likewise, a given template can be assigned to many different profiles.
A signal from a browser to a server that asks the server to send a specific file to the browser. The request, plus some details about the server’s response to the request, is recorded as a line in a log file. Although “GET” in a log file is usually thought of as a “request,” both “POST” and “GET” methods are requests.
With respect to Internet Protocol addresses, indicates success in identifying and displaying a text domain name for a numeric IP address.
How well a site draws visitors back for more visits. Alternatively, a measure of the effectiveness of a source of visitors (a campaign, a search engine, individual keywords on a search engine, an affiliate site, etc.) measured in terms of Recency and Frequency of visitors who were originally introduced to the site by that source.
130 WebTrends Marketing Lab 2
A code in the “status” field of a log file that identifies the success, failure, and other characteristics of a transfer of data from a server to a browser. Also called Status Code. See Status Code page 134 entry for a full list of all error codes.
Visitors who have been to your site before. Returning visitors are identifiable only on sites that give out persistent cookies. WebTrends identifies visitors as returning visitors if they have a cookie from your site dating from before their first visit during the reporting period.
A path that ends at a designated page, called the destination page in WebTrends reports. Reverse indicates “backing up” from a certain page to examine how visitors arrived there.
A group of measures, made up of Recency, Frequency, and Monetary Value, which are useful for segmenting customers for marketing purposes. RFM analysis is a marketing technique used to determine quantitatively which customers are the best ones by examining how recently a customer has purchased (recency), how often they purchased (frequency), and how much the customer spent (monetary value). Requires use of persistent cookies and Visitor History. If visitors did not visit the site during the report time period, their RFM is not included.
A series of two or more pages on a web site that can be treated as a kind of process or logical sequence, such as the process of making a purchase (the checkout process), the process of signing up for a newsletter (the signup or registration process), the process of using a gift finder, and so on. While a scenario by definition has a series of ordered steps, it is possible for visitors to start processes midscenario, such as a campaign that directs visitors to step 2 of the scenario. New scenario visualization capabilities show visitor progress through scenarios, as well as the origin of visits entering scenarios midway and where visitors went after leaving the scenario. Scenarios are defined by the WebTrends administrator.
A report showing the amount of activity at each step of a defined scenario, plus conversion rates for each transition from step to step as well as for the whole process. Examples of scenarios are checkout, registration, or application sequences. New scenario visualization capabilities show visitor progress through scenarios, as well as the origin of visits entering scenarios midway and where visitors went after leaving the scenario.
Scenario Conversion Rate
The percentage of scenarios completed in relation to those that were started.
A simple programming language used to execute tasks. Scripts are often used for pages on the Internet to serve dynamic content and to tailor pages for individual visitors.
See SmartSource Data Collector on page 133.
Search Engine Keywords
A single word within a search phrase, or a search word used by itself. In the phrase “cordless phone” the individual keywords are “cordless” and “phone.” Also called “search keyword.”
Search Engine Phrase
All the words used in a search. In the phrase “cordless phone” the phrase is “cordless phone,” and in the search “phone” the phrase is “phone.” Also called “search phrase.”
A web site that enables users to search for web pages throughout the Internet by entering keywords.
Search Engine Marketing
The art and science of increasing a web site’s visibility and traffic by being listed favorably on search engines for a defined set of keywords and phrases through paid and optimization tactics.
Search Engine Optimization
The art and science of optimizing your web site to improve the “natural” listing or ranking your site receives from search engines for certain keywords and phrases. Often referred to as SEO.
A computer that stores a web site and interacts with browsers to send (“serve”) web pages and other files associated with the web site.
A server error occurs at the web server and receives an error code in the 500 range. Below are examples of some of the most commonly experienced server errors: • • • • • • 500 – Internal Server Error 501 – Not Implemented 502 – Bad Gateway 503 – Service Unavailable 504 – Gateway Time-out 505 – HTTP Version Not Supported
Session, Sessionize, Sessionization
The process of dividing and ordering a list of page views and events in a site’s log into visits or sessions, where each visit includes the sequence of pages viewed by a visitor during a specified time period.
Session Termination Time Frame
The amount of time in which a visitor was not active and thereby defines the end of a visitor session. The default idle-time limit for WebTrends is 30 minutes. In this case, 30 minutes of inactivity by the user causes the visitor session to be counted separately from the next visitor session–when the visitor becomes active again.
A part of a shopping web site where visitors can park items they have selected, presumably for eventual purchase.
Single Access Page
In WebTrends 6.x and before, a visit that consists of only one page view. In WebTrends 7.x and after, these are called “Single-page Visits.”
A visit that consists of only one page view. In Single-page Visits, the page viewed is counted in at least three WebTrends reports: Single-page Visits, Entry Pages, and Exit Pages.
A trademarked technology from WebTrends. SmartSource Data Management offers an alternative to traditional web server log file analysis, collecting information directly from the visitor’s browser (the client) rather than from server log files, improving data accuracy. Special script in a page’s source code is used to transmit page-level data, not “hit-level” data, to a data collection server—dramatically reducing data volume and decreasing processing time. Advantages of using SmartSource include capturing page views resulting from back button use, views of cached pages, and the opportunity to collect extra, customized data not included in normal web server log files.
132 WebTrends Marketing Lab 2
SmartSource Data Collector (SDC)
A specialized web server application, proprietary to WebTrends that acts as the recipient and organizer of data transmitted from web pages by WebTrends SmartSource Tags. The SmartSource Data Collector also validates and generates cookies as part of the data collection process.
WebTrends SmartSource SDC tagging is often used to insert valuable reporting information into the query string of URLs. This is done through SmartSource Parameters, which consist of name-value pairs.
An automated program that crawls widely through the Internet and collects and indexes information, usually on behalf of a search engine or a monitoring company. A spider can often by identified through the User Agent field of a log file, or through its Internet Protocol address.
Faking the sending address of a transmission in order to gain illegal entry into a secure system.
A code in the “status” field of a log file that identifies the success, failure, and other characteristics of a transfer of data from a server to a browser. Also called Return Code. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 100 = Success: Continue 101 = Success: Switching Protocols 200 = Success: OK 201 = Success: Created 202 = Success: Accepted 203 = Success: Non-Authoritative Information 204 = Success: No Content 205 = Success: Reset Content 206 = Success: Partial Content 300 = Success: Multiple Choices 301 = Success: Moved Permanently 302 = Success: Found 303 = Success: See Other 304 = Success: Not Modified 305 = Success: Use Proxy 307 = Success: Temporary Redirect 400 = Failed: Bad Request 401 = Failed: Unauthorized 402 = Failed: Payment Required 403 = Failed: Forbidden 404 = Failed: Not Found 405 = Failed: Method Not Allowed 406 = Failed: Not Acceptable 407 = Failed: Proxy Authentication Required 408 = Failed: Request Time-out 409 = Failed: Conflict 410 = Failed: Gone 411 = Failed: Length Required 412 = Failed: Precondition Failed 413 = Failed: Request Entity Too Large 414 = Failed: Request-URI Too Large 415 = Failed: Unsupported Media Type 416 = Failed: Requested range not satisfiable 417 = Failed: Expectation Failed 500 = Failed: Internal Server Error 501 = Failed: Not Implemented 502 = Failed: Bad Gateway 503 = Failed: Service Unavailable 504 = Failed: Gateway Time-out 505 = Failed: HTTP Version Not Supported
The part of a dynamic URL that is the template. It is usually the part of the URL before the question mark that separates the template from the parameters. Same as URL Stem Field.
In Path Analysis, each page view in the path is a step. In Scenario Analysis, each page in the scenario is a step.
In WebTrends report tables, this usually refers to the total for just the line items appearing in the part of the table on one report page, i.e., that can be seen by scrolling but not by clicking on a “forward” or “back” button. If a table spans several pages, each page’s portion of the table will have its own subtotal. Statistics for parts of the table not shown on the current page will appear as “Other.”
134 WebTrends Marketing Lab 2
Suffix (Domain Name)
The three digit suffix of a domain name can be used to identify the type of organization to which the web site belongs. For example, the suffix .edu implies that the organization associated with the site is an educational organization.
In WebTrends, a matrix or tabular array of results. Each report usually contains one or more graphs and a table. A table may be broken up to span several pages, or it may fit on one page.
When a redirect page is used, the target page is the page to which the visitor’s browser is sent. The term can also refer to the web page that is the destination of a hyperlink.
A collection of WebTrends settings that has a unique name and defines the content and appearance (language, style) of reports to which it is applied. When specified in conjunction with a profile, it determines a complete report configuration that can then be analyzed. In many cases, a given template can be applied to any profile, and a given profile can have many templates. A template allows you to automate and easily customize the content on WebTrends Analytics Reports for a specific business function or user. Templates give administrators and users the ability to customize their views, as well as assign dashboards, reports and language preferences to a given template.
A cookie that is created by a web site other than the one you are currently visiting.
Time to Serve
The time it takes to serve up a Web page to a visitor, measured in milliseconds.
The pages from which most users enter the site or leave the site. Can be distorted by non-human traffic (for example, spiders and robots). Useful to see if lots of people are following a particular link out of the site or whether visitors appear to have a bookmarked page other than the home page.
The suffix of a domain name. A top-level domain can identify a type of organization such as .com, .edu, or .gov, or it can be a country code such as .uk, .de, or .jp. The top-level domain can be used to identify the type of web site.
In general terms, the number of visits, visitors, or activity on a web site.
Comma separated value files (.csv) used to convert analysis information into more helpful report data. Their uses include creating more readable reports and providing drilldown analysis for campaigns and products. They can translate a captured value into another single value or, when using drilldown capabilities, into multiple values that all pertain to the original value.
Number of unique individuals who visited your site during the report period, as identified by a persistent cookie. If someone visits more than once during the report period, they are counted only as one unique visitor. Unique visitors may not perfectly match the number of unique individuals visiting the site, because someone may visit a site from more than one computer and have a different cookie at each computer, or people may share the same computer to access the same web site.
“Unknown” is a possible line item in several WebTrends reports. In geography-related and organization-related reports, “unknown origin” means WebTrends was unsuccessful in looking up an Internet Protocol address or domain name. In first-time versus repeat visitor and buyer reports, it refers to visitors whose browsers did not accept cookies. In repeat visitor reports where all visitors appear as unknown, then the site does not issue persistent cookies.
Uniform Resource Locator. It is a means of identifying an exact location on the Internet. For example, http://www.webtrends.com/html/info/default.htm is the URL which defines the location of the page Default.htm in the /html/info/ directory on the WebTrends web site. As the previous example shows, a URL consists of four parts: Protocol Type (HTTP), Machine Name (webtrends.com), Directory Path (/ html/info/), and File Name (default.htm).
URL Query String
The portion of the URL that contains query parameters.
URL Stem Field
The part of a dynamic URL that is the template. It is usually the part of the URL before the question mark that separates the template from the parameters. Same as Stem.
Portion of a log file that identifies the browser and platform used by a visitor. Also identified through Tags.
All the activity, of one visitor’s browser to a web site, within certain time constraints. A visit is a series of page views, beginning when a visitor’s browser requests the first page from the server, and ending when the visitor leaves the site or remains idle beyond the idle-time limit.
A person at a computer using a browser to visit a web site. A visitor may make more than one visit during a given time period. Note the combination of person, computer, and browser. Since a person may use different computers or even use different browsers on the same computer, it is possible for him/her to appear as more than one visitor because the chief means of distinguishing a visitor is through a persistent cookie or, less desirably, the combination of Internet Protocol address and platform/browser details.
Visitor History is a feature in WebTrends, which when activated, records specific information about the history of your visitors including how often they have visited your site (frequency), how recently they’ve visited (recency), the number of days between their visits (latency), the value of all their purchases (lifetime value), the campaign that generated their first visit to your site, the search engine phrase used most recently to visit your site, and much, much more. Many reports depend on Visitor History being activated, such as any of the Buyers by reports.
136 WebTrends Marketing Lab 2
The Visitor History table has four categories of information it captures, each of which offers a variety of different measurements and possible report combinations that allow visitor segmentation, including: Visit Attributes, Campaign Attributes, Purchase History, and Visitor “Firsts.” Also, Purchase History can measure any form of conversion the WebTrends administrator defines, not just sales. Persistent cookies are used to recognize unique visitors and to record Visitor History events, which are only associated with this unique ID—not specific, known individuals. With all Visitor History measures and reports, a visitor must have visited the site during the report time period in order for their Visitor History data (data which may be outside the report time period) to be included in the report.
A full time period a visitor spends at a particular site. As soon as there is 30 minutes (definable within WebTrends) of inactivity, the session is closed.
Wireless Application Protocol.
A program used on a WAP device to display site content, similar to Netscape or Internet Explorer on PCs.
A server that acts as an intermediary and relays requests from visitors with WAP devices to your site.
A wireless device using Wireless Application Protocol (WAP), such as a cellular telephone or radio transceiver, that can be used to access the Internet. WebTrends reports only include WAP devices if the web data activity file shows the device used a WAP browser.
WebTrends Data Warehouse
The WebTrends Data Warehouse (formerly called the Webhouse Builder) transforms raw web data activity files into a normalized format which can later be used by web traffic analysis profiles for analysis and reporting. Without the WebTrends Data Warehouse, large logs files must typically be stored on a separate machine accessed through a mapped drive, which makes the speed of the analysis dependent on the speed of the network connection. Additionally raw web data activity files are just that, unprocessed, and in their original state. Web data activity files that have been imported and stored using the WebTrends Data Warehouse have already been parsed, normalized, processed, and possibly even filtered, making reporting time for large logs significantly shorter.
Specially named URL parameters that work specifically with the WebTrends Auto-configuration feature. These parameters are created and transmitted by SmartSource Tags or using WebTrends Script, and are recognized by WebTrends to allow automatic generation of reports based on those parameters, without the need for configuration on the part of the WebTrends administrator. For example, parameters can be used to assign a page to certain Content Groups, Scenarios, or to insert data into Visitor History Tables as “first campaign” or other attributes.
Acronym for Wireless Transport Layer Security protocol, which is the security layer endorsed by the WAP Forum (www.wapforum.org). Its primary goal is to provide privacy, data integrity, and authentication for WAP applications.
A visit that included no page views. This is possible if a visit consisted of at least one request for a nonpage file (such as a graphic), but no page files (such as .htm, .asp, .jsp, or .cfm).
A/B testing 7 abandonment rate 117 Accessed File Types report 63 acknowledgement page 117 acquisition 117 email marketing 83 referrers 76 acquisition metrics 73 acquisition metrics worksheet 86 Activity by Referring Site report 77 Activity by Search Engine report 83 activity, web 117 ad 117 ad campaign 117 Ad Click 51, 117 Ad View 51, 117 address filter 61 Address, web 117 Advertising Views 51 aggregate 117 authenticated user 118 authenticated user name identifying visitors 39 authenticated username filter 64 authentication 118 average frequency 118 average latency 118 average lifetime value 118 average recency 118 average, statistical term 118
campaign 118 campaign creative 119 campaign drilldown 119 campaign filter 67 campaign ID 119 Campaign Performance Dashboard 74 campaign type 119 Campaigns report 85 checkout page 119 child profile 119 click 119 clickstream analysis 88 click-through-rate 119 client 119 client errors 119 client-side data collection 120 client-side tagging 21 benefits 22 drawbacks 23 combined log file format 120 commerce web sites 12, 15 complete path 89 consulting with WebTrends 5 content group 120 content group path 89 content groups 46 Content Groups report 47 content path 120 content web sites 14 conversion metrics 87 cost 87 conversion, conversion rate 120 cookie expiration 37 cookie filter 60 cookies 35, 120 first-party 37, 38 problems 37 third-party 37 corporate portal web site 13 cost of conversion metrics 87 critical metrics 11 CRM 108 custom reports 68 Custom segmentation 104 custom segmentation 104 Customer Center vi customer databases 108
banner, banner ad 118 behavior segmentation 99 bookmark 118 branding web sites 18 browser 118 browser filter 60 bugs submitting to WebTrends v business goals 11 business metrics 13
customer relationship management 108 customer retention 10 customer self-service web site 13
feedback, sending to WebTrends v file 122 types 54, 122 file filter 62 filtering data 57 filtering worksheet 71 filters 123 address 61 authenticated username 64 browser 60 campaign 67 cookie 60 custom report 57 day of the week 64 directory 63 entry page 65 exclude 57 file 62 hit 58 hour of the day 64 HTTP method 60 include 57 multi-homed domain 60 Onsite Advertising 63 referrer 66 requested URL 59 return codes 61 visit 59 first-party cookie 123 first-party cookies 37, 38 first-time buyer 123 first-time vs repeat visitors 87 focused path 89 forms 123 frequency 100, 123 FTP 123 funnel 123
dashboard 121 data collection methods 21 choosing 28 data collection worksheet 29 data exploration 107 data farming 107 data integration 107 data record, sample 32 data tagging 21 benefits 22 drawbacks 23 day of the week filter 64 dead-end paths 93 defining behaviors worksheet 56 demographic data 108 demographic data, reporting 94 destination page 121 dimension 121 direct traffic 77, 121 directory 121 directory filter 63 DNS Lookup 121 documents 121 domain names 121 problems 34 visitor identification 34 domain type 122 drill down 122 drill down capability 111 dynamic page 122 dynamic pages URL rebuilding 53 dynamic web page 44
email campaigns, tracking multiple 85 email marketing and acquisition 83 embedded IDs 39 entertainment web site 12 entry file 122 entry page filter 65 entry pages 74, 122 Entry Pages report 75 Excel 107 Excel’s PivotTable function 107 exclude filters 57 exit pages 92, 122 Exit Pages report 93 exit ratio analysis 92 external databases 108
geography drilldown 123 GeoTrends 108 GeoTrends database 123 GIF file 123
hidden referrer 123 hit 124 defined 32 hit filter criteria 59 hit filters 58 Hits Trend report 64 home page 54, 124 home page URL 124 hosted solutions 28 hour of the day filter 64 HTML 124 HTTP 124
HTTP methods filter 60
multi-homed web server 126 multi-page visit 126
identifying visitors 31 idle-time limit 124 include filters 57 informational web site 12 instrumented web page 124 internal search 92 international leads, distribute 8 Intranet web sites 13, 17 IP addresses 124 problems 34 visitor identification 34
navigation 126 navigation measurement 88 new visitor 126 New vs. Returning Visitors report 87 newsletter sign up 9 no referrer 77, 127 non-hosted solutions 28
objectives-critical metrics worksheet 20 Onsite Ad Impressions report 52 Onsite Advertising filter 63 order 127 quantity 127 value 127 order count 127 other, report term 127
landing pages 74, 125 latency 100, 125 lead-generation web sites 12, 16 lifetime value 101, 125 link 125 log entry, explained 25 log file URL 125 log files 23, 125 access 26 benefits 27 drawbacks 27 format 24 rotation 25 loyal visitor 125 LTV 125
page 127 page view 31, 127 paid search phrase 127 palm browser 127 palm device 127 parameter 128 parent-child profiles 69, 128 path 128 path analysis 88, 128 path of interest 128 percent change 128 performance dashboard 74 persistent cookies 36, 128 PivotTable function (Excel) 107 platform 128 portal web site 12 product 128 product groups 47, 128 Product report 48 profiles 129 definition 58 protocol 129 proxy server buffers 35 psychographics 129 purchase 129 purchase conversion funnel 129 Purchase Conversion Funnel report 50, 92
marketing campaign 125 mean, statistical term 125 Measurable Improvement Cycle 5 measures 126 media web site 12 median, statistical term 126 metrics acquisition 73 conversion 87 Microsoft Excel 107 mode 126 monetary value 126 most recent campaign 126 Most Recent Search Phrases report 83 multi-homed domain 126 multi-homed domain filter 60 multi-homed log file 126
140 WebTrends Marketing Lab 2
query parameter 129 query string 129
recency 100, 129 redirect page 129 referrer 130 referrer filter 66 referring domain 130 referring site, domain, URL 77 referring URLs 130 and acquisition 76 registration conversion funnel 130 registration information, reporting 94 repeat buyers 130 report period, reporting period 130 report templates 130 reports 130 Accessed File Types 63 Activity by Referring Site 77 Activity by Search Engine 83 Campaigns 85 Content Groups 47 Entry Pages 75 Exit Pages 93 Hits Trend 64 Most Recent Search Phrases 83 New vs. Returning Visitors 87 Onsite Ad Impressions 52 Product 48 Purchase Conversion Funnel 50, 92 request 130 requested URL filter 59 resellers, finding 8 resolve 130 retention 130 retention metrics 99 retention worksheet 106 return code 131 return code filter 61 returning visitors 131 reverse path 131 RFM 131 rotation of log files 25
server 132 server errors 132 session cookies 36 session ID 39 Session Termination Time Frame 132 session, sessionize 132 sessionizing visits 32 sessions 32 shared key between two databases 109 shopping cart 132 process 90 scenario analysis 90 single access page 132 single jump analysis 90 single-page visit 132 site objectives 11 site structure issues 52 SmartReports 110 SmartSource 132 tagging 21, 133 SmartSource Data Collector (SDC) 133 and cookies 36 and URL classification 45 SmartSource Parameter 133 SmartView 95 software solutions 28 spider programs 133 spoofing 133 static web page 44 status code 134 stem 134 step (in a path) 134 subtotal 134 suffix (domain name) 135
table 135 tag 135 tagging 21 benefits 22 drawbacks 23 target page 135 template 135 third-party cookie 135 third-party cookies 37 time stamp 33 time to serve 135 top pages 135 top-level domain 135 traffic 135 training with WebTrends 5 translation files 135
scenario 131 Scenario Analysis 49, 90 scenario analysis 131 scenario conversion rate 92, 131 scope of analysis, focusing 44 script 131 SDC 133 tags 21 search engine 9, 132 analysis 9 keywords 131 marketing 132 search engine optimization (SEO) 132 search engine phrase 131 segmentation 99 self-referring URLs 77 self-service web sites 17
unique visitors 32, 105, 136 unknown 136 URL 136 URL classification 44 Advertising Views 51 and (SDC) 45
content groups 46 example 45 product groups 47 scenario analysis 49 WebTrends methods 45 URL format 44 URL query string 136 URL rebuilding 53 URL stem field 136 user agent 136
VBScript tag 136 visit 136 visit characterization worksheet 97 visit filter criteria 65 visit filters 59 visit, defined 31 visitor 136 behavior 43 defined 31 goals 11 identification 31 identifiers 33 segmentation 99 visitor history 101, 136 visitor ID worksheet 41 visitor session 137 visitor summary 105 visitors worksheet 115 visit-to-exit ratio 93
self-service oriented 17 strategy 4 structure issues 52 web-customer intelligence 3 WebTrends Analytics On Demand 28 WebTrends Analytics Software 28 WebTrends consulting and training 5 WebTrends Data Warehouse 137 WebTrends GeoTrends 108 WebTrends Marketing Lab 18, 104 WebTrends SmartReports 110 WebTrends SmartSource Data Collector (SDC) 21 WebTrends SmartView 95 well-known parameter 137 worksheet acquisition metrics 86 data collection 29 defining behaviors 56 filtering 71 objectives and critical metrics 20 retention 106 visit characterization 97 visitor ID 41 visitors 115 web log 20 WTLS 137
zero-page visit 137
WAP 137 WAP browser 137 WAP carrier 137 WAP device 137 warehouse reporting 110 web activity 117 collection methods 21 defining 31 web address 117 web analysis focus 11 web analysis introduction 2 web data warehouse reporting 110 web log worksheet 20 web page, dynamic 44 web server log files 23 web site branding oriented 18 business metrics 13 business models 14 commerce oriented 15 content oriented 14 goals 7 intranet oriented 17 lead-generation oriented 16 objectives 11 objectives and critical metrics worksheet 20
142 WebTrends Marketing Lab 2