In the second part of our blog series, we looked at relative criteria and how it’s used to calculate Spotter’s influence score. To come full circle, we’re now going to deal with absolute criteria and illustrate the roles of Audience and Credibility. Once familiar with all 4 pillars we’ll tell you how we combine them to achieve a comprehensive influence score.
Audience is scaled to an influencer’s number of followers, and is fundamental to our score. When I first signed up to Twitter, I used to gauge the authority of Twitterers by looking up their number of followers and following. Although it may seem useful and convenient this method can often be deceitful for community managers and others working with social media. Like most social networks, Twitter is a victim of misuse by people who want to bend the rules to get forward.
As a social media expert it’s important to understand the reason why misleading audience numbers can impact the analysis of your target audience. For example, an account with 50K followers of which 90% is bought, doesn’t give your brand true visibility since most of the followers are inactive or fake accounts.
Sites such as Twitblock or Fake Follower Check are handy free services to find out how many fake followers an account has. These sites however, are unable to provide insights on unusual behaviour in your fields of interest. To identify fake or inactive accounts within the sphere of your brand or business, you need a contextualised influencer analysis solution that detects accounts with an extraordinarily high number of followers vis à vis the times they’ve been listed. (fig. 1)
Deeper insights on who does and does not impact your brand can only be gathered when measuring influence with relevance. For example, the value of accounts abusing the use of hashtags by participating in almost every trending topic on the Twittersphere should not be overestimated when they appear in your domains of interest. Due to their non-channelised interest, these accounts are rarely listed and can be considered as unreliable.
It's Not How Many Followers You Have That Counts
If you want to be accurate in contextualised influencer analysis, you need to go beyond the measurement of followers and look into domain specific interactions as explained in our previous blog post. On some occasions however, even this isn’t sufficient to discern an influencer from another. For such cases the Credibility pillar kicks in and is used as a decisive factor to measure the authority of the influencer.
Credibility is scaled to the number of times an influencer is added to a list on Twitter. Since listing still remains a manual procedure for Twitterers, we consider that it has extra added value, as people only take time to do it when they really feel it is useful to them. This metric allows us to detect unusual behaviour and differentiate bots or publicity accounts from real influencers. This metric is key to identifying misleading accounts that probably have no “real” impact to your business.
Combining the 4 pillars
At this point you would have realised that there’s no single metric to measure contextualised influence. For example, if you scored influence only with Activity, bots would rank as top influencers. Similarly, if you measured influence only with Resonance, occasional tweets by celebrities would rank them at the top of your list. To ensure we provide our clients with a relevant and comprehensive score we’ve merged absolute and relative criteria.
When combining the four pillars, we provide a customisable feature of setting your score for different usages. For example if you were tracking a Live event, the balance of your score would be tipped towards Activity, giving you a score that is sensitive to any changes in trends. Tune it right and you’ll be engaging with the people who matter the most to your business.
To conclude our series on measuring influence with relevance, in the next blog post we will take you through a concrete example of how we gather, measure and analyse the data before scoring it.
Copyright © Akio