How Hummingbird frees content producers from the shackles of keywords

An iCrossing point of view

Prepared by:

Modestos Siotos Natural Search Strategist

This document contains confidential and proprietary information of iCrossing Inc.

How Hummingbird frees content producers from the shackles of keywords
Hummingbird, the latest Google update which was quietly rolled out some time around late August to early September 2013, went unnoticed up until Google publicly announced it one month later. However, this does not mean it’s not a significant update and in fact there is still a lot of confusion and speculation about what it is and how it changes Google’s organic search landscape. The present document outlines the main characteristics of Google’s most significant update in over a decade, it explains why it will change the way people search and suggests some actions online businesses need to make so their content strategies benefit from the latest update.

Quick facts
         Biggest algorithm change since 2001 Biggest infrastructure change since Caffeine update (2010) Affects 90% of search queries More intelligent understanding of search queries Understands query intent based on concepts, not just keywords Answers queries made in natural language Conversational (adaptive) search based on previous query Does not filter or penalise websites Higher impacts on long-tail queries

What is Hummingbird?
In a nutshell, Hummingbird improves Google’s perception of queries that use natural language. With the ever increasing adoption of smartphone and tablet devices the way people search is changing. The keyboard as an input device isn’t ideal for people on the move where vo ice commands are more natural and quicker to make. Google had to adapt to these trends in order to meet these users’ needs as well as to quickly give them the most accurate results.

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

1

With Hummingbird Google is moving away from the traditional keyword-based search and stepping into natural language search, including conversational search. This seems to be the most revolutionary change that has taken place in the industry in several years and one that is expected to significantly change online search. However, in order for Google to turn this into a reality they had to overcome some serious challenges such as understanding the context and user intent behind a given search query. There are several signals Google are currently using in order to work out the user intent behind a search query including:     Users’ physical location Users’ type of device Users’ past searches and social interactions Local time

The above are then filtered through Google’s Proprietary knowledge base to verify the context.

Relationship with Panda & Penguin
Hummingbird is a major infrastructure update that should not be confused with other high-impact updates such as Panda and Penguin, both of which act as additional filters on Google’s main index. Unlike Panda and Penguin, Hummingbird does not filter or penalise pages (or websites) because it only affects what Google deems relevant for a given search query. It does, however, build on these updates along with Caffeine, all of which have been introduced over the past three years:    Caffeine –Significantly improved crawling and indexing times Panda - Lowers rankings for sites with low quality content Penguin – Lowers rankings for sites with unnatural links

The following timeline illustrates all major milestones in Google search to date that progressively led to its most revolutionary update; Hummingbird.

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

2

How does Hummingbird affect SEO?
It may take some time for users to have confidence in the fact that queries made in natural language will actually return the best possible answers. In any case, the shift in searchers’ behaviour doesn’t mean that the content optimisation in the new (post-Hummingbird) search era has to be radically changed over-night. Hummingbird, unlike Panda and Penguin, is not an algorithm update the SEO industry needs to instantly adapt to in order to re-align with the latest Google trends. Instead, it is an update that sees Google trying to bridge the gap between the actual user intent expressed in search queries using natural language and what is being retrieved as the most relevant search results. Most of the 200+ different signals previously used by Google to rank pages are still here and unlikely to go away any time soon. Content relevance, the link graph, authority, trust, as well as user engagement signals are still the main ranking influencers as in the pre Hummingbird era. However, it can be argued that Hummingbird, over time, will allow Google to better process signals from social media platforms so in the near future these can become a bigger ranking factor and act as proxies to verify the link graph.

Hummingbird isn’t about more content
Improving onsite relevance is no longer about more keyword-heavy content but about more unique, engaging, authoritative and trusted content which will be socially shared and at times linked to. This type of content will not only be visible for generic keywords but also for long-tail ones as long as Google is confident that it perfectly meets the intent of the user query. Previously, Google’s informational retrieval process was over -focusing on the existence of keywords found on a page. With Hummingbird it tries to step away from this easily gamed not user-focused approach. It is now possible that the copy of the highest ranking page may include none of the keywords that formed the original search query. In this respect, Hummingbird is about better content and not about more (optimised) content. In more technical terms Google’s informational retrieval process is now more user-focused than ever before, aiming at pages that match the user intent and not necessarily the queried keywords. The search giant is now able to understand the relationship between different entities using data from synonymous and substitute queries. With Hummingbird, Google seems to have placed more importance on trending, fresh and trusted content. For content publishers the way forward is being seen as a quality resource and this can only be achieved by constantly covering the latest topics as well as by investing in great authors with a strong reputation for digital copywriting (authorship).

Moving from keyword research to topical research
One thing that will certainly change is the overreliance on and repetition of targeted keywords on “optimised” pages. Copywriters finally have more freedom to produce content that focuses on increasing user engagement rather than on how to please the search engines. Using sophisticated adaptive algorithms, Google is constantly learning from the content patterns it comes across. This is why content producers are no longer limited to a narrow selection of terms with high commercial value. Instead, they can (and should) now use a broader selection of terms, which are tailored to consumers rather than the search engines. Trying to address every single possible question a user may have about a product or a service with content isn’t the way to benefit from Hummingbird’s enhanced understanding of natural language. Similarly, producing content to address every single keyword combination will not help either. The biggest opportunity in terms of content production is offering the best possible user experience, which most of the time depends on the usefulness, quality and presentation of the content. New pages should not be produced with the sole objective of matching a few targeted keywords but with the aim to cover new topics in such a way that users are more likely to be interested in reading, engaging with and sharing a piece of content. This type of content is more likely to match the user intent behind complex queries, which previously the search engine would have considered as vague or ambiguous.

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

3

Therefore, any content strategy should now be led by topical (or conceptual) rather than keyword research. In other words, content producers should avoid blindly mapping keywords into pages without thinking about the actual target audience. Persona profiling based on user data and demographics should always be the starting point, so the targeted audience is identified and taken into account at the very early stages of the content production process. This will require a radical shift not only in the content producers’ mind-set but also all the people who influence them (like SEOs). For many organisations this will be a challenge as this type of change cannot happen overnight.

Why Hummingbird justifies organic keywords data loss
From our point of view, it is therefore hardly surprising that the launch of Hummingbird coincided with the disappearance of organic keywords data from Google analytics. The following example will help you understand why Hummingbird and keyword data loss may have taken place at around the same time. Voice language search in general generates longer queries because it’s based on natural language. On the other hand, typed search queries are much shorter. For instance, here are typical voice queries people walking on the streets of Brighton looking for a nice pizza may make: 1. 2. 3. 4. 5. Where is the best pizza place around here? Show me the best pizza place in Brighton Help me find the best pizza place in Central Brighton Where is the greatest pizza restaurant in Brighton? Find me an award winning pizza place in Brighton

With Hummingbird, Google can understand that the above five long tail queries in natural language have pretty much the same meaning and user intent as the much shorter typed query “best pizza in Brighton”. In this case all five long tail queries are very likely to be combined and translated into the shorter typed one. Considering the usefulness of all the above key-phrases appearing in analytics things may get a bit tricky. Trying to create new or optimise existing content to accommodate the above key-phrases e.g. by reproducing the above questions as well as answering them, isn’t what Hummingbird is about. The latest update will, instead, combine the five key-phrases into one and return the same exact page to all users. In this example, organic keyword data if misinterpreted and can mislead rather than aid the content strategy. Similarly, consider the following voice search conversational queries: “Show me images of Big Ben”

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

4

“How tall is it?” Note how Google “remembers” the subject matter ( Big Ben) and instantly answers the question utilising data from the Knowledge Graph.

“Who designed it?”

“How can I get from Brighton to Big Ben?”

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

5

Now think about a data analyst that comes across the above key-phrases in analytics. In this case knowing that the site’s Big Ben landing page has received more visits compared to a year or month ago is a more valuable insight than focusing on vague keywords such as “how tall is it” and “who designed it”. Therefore, coming up with all sorts of questions and trying to answer them with new content just for the sake of traffic is a rather short-sighted content strategy approach. On the other hand, producing useful content, which is tailored to a very specific audience that gets discussed and shared will almost certainly have bigger long-term benefits. And there is little doubt that by removing organic keyword data from analytics Google is hoping to speed-up the mind-set change of all digital marketers that still think keywords, rather than topics are what matter. Switching off most (or all) keyword data is a way for Google to enforce this change in attitude.

What’s next?
Currently, Google answers some of the searchers questions directly from within the search results if relevant data is available in its knowledge graph. However, the data comes from limited sources (e.g. Freebase, Wikipedia) and ultimately Google is aiming to instantly provide answers by combining information that appears in trusted sources in the web. So, instead of serving pages that are likely to answer a user query, Google will provide the answer directly in the search results. This will almost likely be the next biggest change in search, as we know it today. For instance, Google will one day be able to instantly figure out the intent of a query like “where is the nearest place for a Sunday roast serving great ales?” and provide the best answer based on the following actions: 1. Detect user’s current location 2. Prepare a list of nearby local pubs 3. Pull out ratings based on user reviews that appear on trusted web sources for those pubs Sunday roasts. 4. Repeat the same action based on reviews for the best ales. 5. Take into account link and social graphs. 6. Sort the results according to the overall score of the above ranking signals. In this example, it is unlikely that any of the sites Google used to provide the best answer to the query will be visited by the user because the answer will instantly appear in Google’s search results. This will have negative implications on the traffic levels of those websites that feed Google with the above information, as they will not get credited by a visit.

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

6

Conclusion
Google can now understand and process queries in a more intelligent way. Heavy focus on specific keywords will become less and less important over time as Google’s understanding of queries relies less on the keywords that comprise a search query and more on semantics. Blindly optimising for long tail keywords in this new era of semantic search would make little sense. By not providing keyword data Google will enforce a radical shift in content strategy, moving from keywords to topics or as they say it “from strings to things”, where “strings” stand for keywords and “things” for entities with attributes.

For further information, or if you have any questions, please do not hesitate to contact your iCrossing account manager or results@icrossing.co.uk

© COPYRIGHT 2013 ICROSSING LTD

NOVEMBER 2013 | How Hummingbird frees content producers from the shackles of keywords

7

Sign up to vote on this title
UsefulNot useful