16 November 2007

Return on influence: metrics and measurement for brands in social media.

One of the emerging dynamics of social media is the compelling need for companies and organisations interested in tracking their brand, positioning, reputation and saliency within this space is to have some way of measuring return on both investment and influence.

Many organisations want to be able to measure their involvement or even assess potential returns prior to investment but are slowly coming to the realisation that traditional web-based analytics alone are not going to do the job, so the next question is crucial – what and how should be measured?

The crux of the issue is that regardless of the success of brand monitoring companies like Neilsen BuzzMetrics and TNS Media Intelligence/Cymfony, at this point in history their work is both highly speculative and also completely unique to the interpretation of the entity doing the measuring.

While the ability to track the tone, perception and nuances of discussions around a company and its brands, either via social media or even within it, cannot be achieved through web metrics or traditional advertisement measurement alone, the real problem is that even among the eight or so market leaders Forrester Research identified last year, there is no universally agreed standard upon which to base either measurement and metrics.

Certainly, the measurement attributes identified by Factiva, one of the vendors included in the Forrester study, could likely be considered to be part of any brand monitoring mix but this mix is also likely to be far more dynamic than anticipated.

Last year Factiva tried to get an understanding of the prevailing thinking, hosting a roundtable of influential bloggers and social media critics which identified the following measurement attributes (I've left out explanations of those that need no explanation):

1. Analytics and activity

2. Community activation/call to action

3. The "Conversation Index", define as the ratio between blog posts and comments-plus-trackbacks. This attribute is one subset of measuring participation and engagement but when I look at the majority of external corporate blogs alongside trackbacks and comments even in our brand sector, the actual ratio is low.

4. Demographic/s

5. Influential ideas or memes which the "intensity" or "velocity" of the spread of an idea or message over time. (A "meme" refers to an idea or discussion that grows and spreads from individual to individual into a lengthy chain of commentary. Rather like an idea train.)

6. Participation and engagement, where a recipient not only responds to a message but acts on it as well.

7. Reach, where it’s a lot less to do with sheer numbers than with influence.

8. Relationships and connections. For example, Factiva’s panel noted that a blogger may have a large, diverse audience but he or she may not have an intimate relationship with specific influential individuals within any given community.

9. Relevance, the question being how relevant to my company is a particular blog post among hundreds of a given meme?

10. Sentiment/tone/favourability

11. Content: the focus of coverage of social media. In most cases, this means a narrow, but extremely detailed focus on one subject or issue and it explains the importance of certain blogs over others within their respective niches.

Factiva’s panel concluded that to be successful companies need to use a combination of several key measurement attributes to understand and drive their own social media success for internal and external stakeholders.

Brand monitoring companies need to be able to help clients develop and execute plans which monitor and measure these changing attributes within a specific context and within a sphere of operations. Brand agencies then need to able to be bipartisan and develop strategy that acknowledges both measurement and metrics are dynamic and that a company's ability to identify those key attributes important to it might actually be limited. Setting these against a wider brand context and against other metrics might be another way of helping to define their brand influence.


glennfan said...

We weren't expecting a magic formula to reveal itself at that event, but we were hoping to start a dialog about the issue. I think we did this. The problems with measuring, say, a brand's footprint on the Web are very different than in the traditional media. One of the challenges I see is that the specifics of the conversation methods are constantly changing. When we starting looking at this we knew, for example, that we'd have to take podcasts into account, however none of us were thinking about Facebook.

Facebook now is an emerging place where corporate reputations can be impacted as groups are formed and people turn to them as indications of a brands quality.

Any sort of systematic measurement needs to be dyanamic enough to take these rapidly changing communication methods into account -- making normalized metrics near impossible.

So, my take is that Web measurement will necessarily stay more fluid that traditional media measurement.

John Cass said...

Part of the reason why I asked my question on linkedin.com (thanks for inviting me to your post here) is because I've started to see that several organizations are using one metric for measuring the level of influence of a blogger – sentiment, though this metric is often combined with whether the writer is also a customer or user of an organization's services. If a customer has negative sentiment an organization in this context is invariably going to respond to that person, no matter what their level of influence in the community. In the organization's eyes a person's influence comes from their relationship to the organization and their sentiment towards the organization or product.



Sentiment seems to me more like intention. How is this being measured? In a brand's verbal identity you can look at aspects such as tone of voice, active/passive language, style and context etc and all these psycholinguistic evaluations would seem to me to be appropriate in this context. As I mentioned, the level and breadth of the "chatter" would also be consistent measurements. I know Neilsen is looking at how to measure the total media consumption of individuals and it would seem to me that we are not far from a time where the adoption of brand analytics could be used to examine an individual's full interaction with a brand. This would go well beyond any short term qualitative analysis.