Misleading Indicators – Followers & Friends

Saw this on my twitter feed yesterday:

What immediately struck me was the implied assumption that the number of followers you have infers a level of influence. In our opinion that’s a risky assumption to make especially if you are going to make a business decision using this as a key metric.

Here is what I sent back to Jeremiah via DM:

Let me translate my Twitterese….

The number of followers is not a direct measure of influence. Too many ‘experts’ in the social media field believe that it is and continue to sell this notion. I can quickly and easily increase the number of my followers using hashtags and keywords that are popular. Yet that doesn’t necessarily mean that I am a stronger influencer than I was with a lower number of followers.

Those folks with a larger number of followers should not necessarily receive special treatment from brands. The number of followers or friends a person has on Twitter or Facebook really has minimal bearing on their actual influence. (I know that’s a bit heretical, but I’ll get to the why in a little bit.)

• How many people have used the various advertised services to build their followers rather than organically growing their followers by posting relevant content and ideas?
• How many people send an invite/friend request/twitter follow to every email address they have expecting the ‘polite’ return linking/friending/following behavior?
• How many of the top people in terms of followers have a large brand behind them, providing follower building support? (Example, if you only tweet about HP or Oreo Cookies you’ll develop following due to the power of the brand not necessarily because you are a thought leader in the space.)

Because these numbers can be manipulated, they are not to be trusted as a direct metrics proxy for influence.

The example that I use in our social media workshops uses a metric that everyone thought was a useful metric way back when in 2000-2003: Hits. The logic at the time was that the more hits there were in a given period of time, the better the site was in meeting its goals. But alas, this metric could be easily manipulated. Want more hits? Add more banner ads, objects, photos, etc. to the page. Voila! Higher counts so more success, right? Well, not really.

Follower counts are the same as hit counts. Look at some of the top people on Twitter with 5,000+ followers. If they are focused on a single topic, they probably do have influence. But most people are not that focused, tweeting about business, sports teams, their family, current events, pets, politics, etc. Do these folks really have a sphere of influence that marketers can embrace and attempt to cultivate through the Twitter Celebrity? Hard to tell.

The idea of identifying influencers in an easy to understand and quick manner has been a search for the holy grail since online communities started becoming more popular in 2000. At Participate.com, we hired smart people to analyze metrics and activities to develop relevant networked connections that indicated a level of influence within the community.  We used the new techniques of social mapping as well as relationship metrics of interactions. The work was never easy and it never gave a true understanding of influence. What did give some insight into influence, was looking to see how others interacted with the individual, not how many individuals read his or her content.

For marketers, a better way to measure influence is to analyze the content being added on Twitter in conjunction with analyzing who the person’s followers are. This is a tough, manual project. But in the end, you’ll have a much better understanding of whether or not a particular individual with a high following is actually an influencer.

As much as we want one, sometimes there is no holy grail. Using simple metrics as proxies is not a substitute for the hard work that data analysis takes to prove a hypothesis. Don’t fall for the trap of taking the easy way out.

Have a different opinion? Please share your thoughts with us in the comments’ section.

3 Responses to “Misleading Indicators – Followers & Friends”

Adrienne says:

December 26, 2009 at 10:36 am

Good points about Twitter. People underestimate the power of bots on Twitter as well.

MRowland says:

February 24, 2010 at 5:01 pm

Now we’re doing an experiment using the available tools out there and the advice from “Twitter Experts” on how to game the system. We’ll be publishing the results on our blog. Our hypothesis is that the new followers are not as valuable as those who find you organically through your content.