By Chris Lee

Digital media writer and consultant Brian Solis argues that popularity does not equals influence when it comes to online marketing. In effect, the number of Twitter followers, Facebook fans and blog readers does not determine an organisation or individual’s influence, more the impact the organisation has on encouraging action among that followership.

Various metrics have been developed to measure Twitter influence, for example, such as Klout and PeerIndex, but measurement is such an important field so how can marketers get it right? To get answers NMK got in touch with Flemming Madsen, founder of online analytics firm Onalytica.

The Challenge

According to Madsen, “not all voices are created equal” when it comes to measuring online influence. Understanding their impact is more important. Madsen argues that traditionally, consumer-focused market research has attributed equal weight or influence to all voices or respondents, because in the past, the assumption that everybody’s response is equally important was accepted when those being influenced by an individual were neighbours, friends, family or work colleagues.

“Today, it’s a whole different ball game,” Madsen said. “Write a review of a car, restaurant or baby food online and you could be impacting millions. The challenge is establishing whom and what really has influence. Influence is a complex and dynamic area. Relying on what we think we know about key influencers is no longer appropriate. People often confuse popularity with influence and one cannot be used to measure the other. Influence is important because when voices mentioning brands are adjusted for influence it is possible to correlate share-of-noise into share-of-market.”

Three levels of influence

Madsen argues that there are three levels of stakeholders and their influence. These include:

• Popularity: Popularity concerns how well-known a source is. As humans we tend to overrate the importance and relevance of those we hear about all the time, and similarly underrate those with whom we are not so familiar

• Influence: We can define influence as the impact a source has on a particular issue, and a source does not have to be popular to be influential. It is who listens to the source that is significant and who they then impact, not how many listeners they have

• Relative Influence: is the influence-to-popularity ratio of a source. The cost of engaging with a particular stakeholder is largely related to their popularity and hence a popular celebrity may significantly increase the cost of a campaign yet yield less success.  On the other hand, it may be easier and more cost effective to engage with a less popular stakeholder with high relative influence – delivering a comparatively higher return on engagement than their popularity would lead us to believe possible

“Sources of influence will vary according to the topic under discussion,” Madsen argued. “For example, someone who is a respected commentator on politics is not necessarily going to have the same degree of influence when talking about knitting. So it is important to make sure that the insight monitoring methodology or service being used gives different weighting according to the context or topic.”

Understanding influence

Madsen said that understanding influence is crucial when using the online debate to predict business outcomes or the results of traditional polls, for example. By analysing changes in awareness and preferences among top influencers in a particular market, it is possible to predict a change of perception and purchasing attitudes among consumers due to the impact these influencers have. If more key influencers are talking about a product, then it follows that there will be an upturn in sales and equally the results of traditional consumer polls can be predicted as public opinions are largely formed by influencers.

So once organisations have got all this information, what do they do with it?

“Our advice is to keep it simple and start with what already exists,” Madsen advised. “We recommend mapping insight monitoring against known key performance indicators (KPIs).  For instance, one of our customers uses its online insight results as direct input into its monthly reporting to senior management, as well as seeing how it correlates against regular voice-of-the-customer surveys.”

Other examples of KPIs include measurement of specific market campaigns, customer satisfaction, and comparison against competitors.  

“Ask other departments what intelligence they would like, but to which they may not currently have access.  After all, there is little use in pointing out market trends to a department if they have no intention of utilising it, or if it does not map on to their specific focus or objectives,” Madsen said.

Madsen also warned that online insight monitoring is not an answer for everything. It may provide a continuous ‘listening in’ of the online conversation, hearing what people are really saying rather than being prompted, but traditional market research is better at measuring the reaction to abstract or hypothetical ideas, or providing intimate, highly-detailed information that interviews provide.

“In reality, traditional research and online monitoring are likely to co-exist for a long time to come and the latter certainly provides the instant finger-on-the-pulse that so many companies want. However, without understanding the true nature of influence, then they may be limiting themselves to a fairly crude tool,” Madsen concluded. “The good news is that understanding influence is now very much possible.”