You are on page 1of 8

Conversation Impact™: Ogilvy’s Simple, Results-Driven

Social Media Measurement Model for Marketers


Irfan A. Kamal
Vice President/Digital Influence Strategist
360° Digital Influence
Ogilvy Public Relations Worldwide
irfan.kamal@ogilvypr.com
John H. Bell
Managing Director/Executive Creative Director
360° Digital Influence
Ogilvy Public Relations Worldwide
john.bell@ogilvypr.com

Track the few metrics


that really matter.
With two-thirds of the world’s Internet population
now visiting a blog or social networking site, driving
results through social media has become
an important component of the marketing strategy.

To help guide brands on social media spending


decisions, Ogilvy’s global social media marketing
group, 360° Digital Influence, has developed
and introduced a new business objective-driven
Contact: model that provides a quantitative measurement
Irfan A. Kamal
framework for [social] media effectiveness—
Vice President/Digital
Influence Strategist Conversation Impact.
360° Digital Influence
Ogilvy Public Relations Worldwide
Tel: 202 729 4273
irfan.kamal@ogilvypr.com

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 1 of 8
Conversation Impact™ (continued)

In developing Conversation Impact,


we had three key goals in mind:
1. The approach must allow for cross-channel
performance comparison, specifically including
the social media “channel”
2. The metrics must include actionable data for
in-market campaign optimization
3. The model must be simple, objective-driven
and cost-effective enough to use for every
campaign, social media and “360”/multi-channel
In addition to introducing the model below, the following sections
walk through some specific considerations and objectives we
worked through.

Measure cross-channel.
De-emphasize metrics that don’t
allow apples-to-apples comparisons.
Current approaches to advertising focus on measurement of such
items as ad recall, ad reach, ad frequency, in-category brand aided/
unaided recall (for evaluating awareness/consideration), intent to
purchase and net promoter surveys (for evaluating preference);
and—for measurable media such as online ads—action/conversion
rates (for evaluating action or conversion).

Traditional metrics are not readily applicable to the analysis of


social media—but are still being used. Reach/exposure (the
number of people exposed to the message) is often used as a
metric related to brand awareness, positioning and preference
campaigns. However, word of mouth WOM is consistently trusted
1
more than other forms of communication/marketing , so direct
CPM or impression-based comparisons are not useful.

Another complicating factor is that simply applying an adjustment


factor to a WOM mention is still not useful because of the varying
1 eMarketer Bridge Ratings and University of Massachusetts 2007

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 2 of 8
Conversation Impact™ (continued)

influence levels of the conveyor of the message. That is, people


trust messages differently based on context (which includes factors
such as the person conveying message, tonality of message and
specific content of the message).

Finally, simply reporting “activity” metrics like page impressions,


interactions and time does not help marketers determine whether or
not a campaign was successful in driving their ultimate marketing
or communication goals. Even measures of “engagement” cannot
connect activity to core marketing goals.
Conversation Impact simplifies and standardizes measurement in a
way that is readily recognizable to marketers, while accommodating
both conventional and new metrics and data that account for these
differences in social media.

Provide actionable measurement


that focuses on key goals.
There is rich data now available on a continuous basis, including daily
data from the semantic analysis of the millions of conversations in
social media. We integrate these analytics into the model to help drive
optimization.

As an example, for a brand positioning campaign, we can evaluate


preference and action in ways that help us understand which social
media influencers are adopting which types of messaging and in
which channels. We look at what people voluntarily say and do
across the social Web. We can use this data to help guide ongoing
creative and spending decisions.

Use a model that’s both strategic


and simple.
We organized our model directly around 3 main goal categories that
build off years of “marketing/communications funnel” research into the
best way to drive action:

• Awareness & Positioning


• Preference
• Action
Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 3 of 8
Conversation Impact™ (continued)

Table 1 presents the key representative measures within each of


these categories.
Table 1: Conversation Impact Model with Representative Metrics

Metric / Funnel Goal > Awareness Consideration Preference Action Loyalty

Reach/Positioning
• UMV – blogs, site, microsite, applications, other, total #/%
change

• Volume of online conversation, #/% change, Cost per Online


Conversation Generated (CPICG)

• Share of voice in category (=Volume for brand/volume for


category). #/% change, Cost per Increase in Share of Voice
(CPISV)

• Search visibility (for relevant keywords)

Preference
• Sentiment index of online conversation (% positive–% nega-
tive), #/pts change, people reached vs. all, Cost Per Increase
in Sentiment Index (CPISI)

• Share of positive voice in category (= brand positive mentions


/ category positive mentions), # / % change, people reached
vs. all, Cost per Increase in Share of Positive Sentiment
(CPISP)2

• Relative net promoter score (NPS) in category (=brand NPS /


category NPS), #/% change, people reached vs. all, Cost per
Point Increase in NPS (CPINP)

Action
• Registration: RPA, CPA, $, #

• Sale: RPA, CPA, $, #

• Advocacy: RPA, CPA, $, #

Some of these metrics are based on data from social media moni-
toring software; other metrics are obtained from server logs, Google
and other analytics data, and surveys.

Here are examples of data sourcing and definitions of key metrics


shown above:

Awareness/Reach/Positioning
• Share of total voice within category = volume of mentions for
brand / total volume of discussion in category3

2 Depending on volume and the prevalence of negative discussion, we may look at


share of net positive voice, where net positive voice = brand positive mentions –
brand negative mentions 
3 The denominator may also be modified to include only the brand plus specific
competitors

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 4 of 8
Conversation Impact™ (continued)

• Calculated via social media monitoring/listening software or


through a direct consumer survey
• Cost per point increase in above metric

Preference
• Share of positive voice within category = volume of positive-
sentiment mentions for brand/total volume of positive-sentiment
mentions in category
• Can also be compared to / calculated as share of net positive
voice, which = volume of (positive – negative) mentions for brand
/ total volume of (positive – negative) mentions in category
• Calculated via social media monitoring/listening software or
through a direct consumer survey
• Cost per point increase in above metric
Action
• Campaign- or influencer-attributable actions
• Calculated using tracking analytics or through Ogilvy’s social
media activation platform technology

Note that the measures shown are representative, not


comprehensive—the key focus of the model is to use categories
and metrics that provide simple, useful data in ways that provide
for ease of comparability and analysis.

Evaluate campaign impact on


influencers, consumers or both.
We used the model to evaluate two example brand campaigns. We
selected these two campaigns to demonstrate the flexibility of the
model in measuring impact on both influencers and consumers.
We define an “influencer”—someone who is effective at broadly
distributing a message or driving action—based on a number of
factors, including a person’s connectedness, reach and ability to
engage and drive results around a specific target audience and
discussion topic. For the purposes of this discussion, “consumer”
represents traditional end users/target audience members.

We employed our social media listening post and survey-based data


collection methods for these campaigns.

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 5 of 8
Conversation Impact™ (continued)

Campaign 1: Consumer Impact—Evaluate Impact


based of a Social Media Community Program on
Consumer Aided Awareness and Preference
Campaign 1’s social media program analyzed consisted of a blog
community-based program. We evaluated pre-campaign and in-
terim measures based on a survey instrument. Measures evaluated
included aided awareness and purchase intent.

Figure 1: Consumer Impact—Awareness and Preference

Social Media Campaign Impact: Consumers


n Pre n Post
67%

48% 48%

20%
10%
6%

Aided Awareness Intent to Purchase (9 or 10) Positive Opinion (9 or 10)

Campaign 2: Influencer Impact—Evaluated Impact


on Influencer Preference for the Brand based on a
Multi-Channel Campaign Including Social Media
Campaign 2’s social media program included social media compo-
nents and a multi-channel traditional and online
advertising campaign.

We used a software monitoring product to track and categorize the


tonality of social media mentions around the brand. This product’s
algorithm classifies social media discussions based on a trainable,
semantic, natural language-based categorization algorithm and
has a high degree of consistency and reliability in assigning social
media mention sentiment and topic relevance4.
4 It should be noted that our team’s approach is software-independent. We set up
Social Media Listening Posts™ with underlying software that is the most appropriate
for the specific client engagement.

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 6 of 8
Conversation Impact™ (continued)

Our focus for this measurement program was to evaluate the


impact on influencers driving brand positioning and preference
in social media; specifically, we evaluated the number of positive
mentions as a share of total positive mentions within the brand’s
competitive set. There were five competitors identified to be part of
the competitive set.

For Campaign 2, we found an increase in influencer preference in


social media mentions—as measured by relative share of positive
voice in social media—of 1.5 percentage points during the first 3
months of the campaign. This represented a 9.2% increase in pref-
erence among those posting in social media about the brand.

Figure 2 shows the monthly trend in preference.


Figure 2: Influencer Impact—Share of Positive Voice in Social Media

Social Media Campaign Impact: Influencers


Share of Positive Voice in Social Media
19.0%

18.2%
18.0% 17.7%

17.0%
17.1%
16.0%
16.2%
4-month Trend
15.0%

Dec-08 Jan-09 Feb-09 Mar-09


Pre-Launch Campaign Launch

Use the model today.


In our early uses of the model, we’re finding that Preference and
Positioning are two key areas in which social media can deliver
particularly efficiently—and we’re now implementing new Action-
based campaigns to further expand the range of social media goals.
Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 7 of 8
Conversation Impact™ (continued)

We believe the model represents a useful step forward in social


media impact measurement, primarily due to its focus on tracking
metrics with comparability across different types of advertising and
communications.

As social media increasingly becomes a standard component of both


advertising and communications campaigns, this type of simple,
cross-channel comparative framework will become more useful in
answering media allocation questions and helping marketers decide
which social media efforts are worthy of scaling further.

Creative Commons with Attribution license. Attribute: Ogilvy 360 Digital Influence. page 8 of 8

You might also like