16 Comments

Summary:

Though navel-gazers may have freaked out this morning after seeing their Twitter follower counts reset to zero, a new paper out of the Max Planck Institute for Software Systems in Germany finds that Twitter follower numbers are a poor indicator of influence.

If you’re the type to measure your self-worth by how many people follow you on Twitter — and perhaps more importantly, who — this morning was rather exciting. Gizmodo published directions for a simple way to make other Twitterers follow your account, resulting in some users gleefully playing with the loophole to make it look as if they were tight with the likes of Ashton Kutcher and even Jesus. Twitter’s clunky way of fixing the problem included temporarily resetting follower counts to zero, which set off additional squeals of protest from from the navel-gazers.

There's little overlap between the different types of influential users on Twitter. (Indegree refers to # of followers.)

But does your number of followers actually matter? A new paper out of the Max Planck Institute for Software Systems in Germany finds that Twitter follower counts are a poor indicator of influence. Interestingly, the research, which was led by a post-doctoral researcher at the institute by the name of Meeyoung Cha, found two quite different forms of engagement on Twitter. First, retweet influence — when a user’s content is likely to be passed along — and second, mention influence — in which users engage one another in conversation by mentioning each other by Twitter handle.

Cha’s study, for which Twitter administrators contributed data, looked at the relative influence of the 6 million active Twitter users it counted within 52 million active accounts circa August 2009. First of all, retweets are about content; 92 percent of them contained a URL. Influential content aggregators include technology-specific sites like Techmeme, but the study also indicated that sources like the New York Times have widespread influence, and are retweeted on a variety of topics. However, it’s possible for relatively unknown users to gain influence and rise from obscurity by focusing on a single topic.

Meanwhile, the most-mentioned users were celebrities. Only 30 percent of all Twitter mentions contained a URL, indicating that they’re more personality- than content-driven. I’d be interested to see a deeper look at the retweeters and mentioners themselves. I wonder if there are two types of Twitter users: those who go on Twitter in order to talk to celebrities (and only rarely get a reply, in many cases) and those who tweet to an audience of their own, sharing content with followers with via retweeting.

Sean "Diddy" Combs is a rare Twitter influential on all three vectors studied.

But back to the Max Planck study: There’s very little overlap between the different types of influence on Twitter. Only two users made the top-20 lists for followers, retweets and mentions: Ashton Kutcher and Sean “Diddy” Combs.

There’s at least one startup, San Francisco-based Klout, that focuses specifically on measuring influence. Klout accounts for more factors than the Max Planck study, including number of followers, retweets and mentions, but also follower-follow ratio, the number of lists a user is counted on, how many unique retweeters they have, how many inbound messages per outbound message, how many updates and many more. Then all those factors are applied to everyone in the network who retweets, mentions, lists and follows any user, and the value of users’ networks are added to their own score. The recently funded Klout makes this data available via an influence score. It has signed its first paying customer, the Idealab company TweetUp.

Related content from GigaOM Pro (sub req’d):

The Dos and Don’ts of Social Media Marketing

  1. Sorry Liz – this is rubbish and here’s why. It sure does matter how many followers you have – forget about rt’s and sharing non-site links. But when Om shares a link or Mashable shares a link – you know, sites that got the free million+ cash-generating Twitter gift, they/you get more traffic.

    So to say it doesn’t matter is just not true.

    It also matters because the more free twitter followers you got in the free gift, the more chance you have to get massive RTs.

    Share
    1. Allen, this research looks at the number of retweets specifically, and doesn’t find they necessarily correlate to accounts with lots of followers. I think you’re talking about the recommended users list of yore, which gave selected users lots of followers but not necessarily people who were interested in them, people who stayed on Twitter, or people who actively tweet.

      Share
  2. I have to say, Twitter counts do matter. It doesn’t matter if they generate genuine conversions. The public perceives you differently if you have more followers, and that is fact.

    That said, if we are talking about conversions only, then followers are meaningless.

    At this point in time, however, follower counts are still the measurement by which a user on Twitter is ranked, and I don’t see that changing anytime soon.

    Share
  3. Liz: You should know there is more than one company out there that measures Influence. Twitalyzer also helps organizations understand their influence and overall impact on Twitter.

    Share
    1. Thanks.

      Share
      1. You’re welcome. By the way, I do agree with the premise of your article and the referenced research paper by the Max Planck Institute. Much our data supports that # of followers does not equate to influence

        Share
  4. Seems like all the gurus are saying followers counts do not matter tonight (http://www.socialnews.biz/tag/Twitter). I agree 100%. I’ve had tweets to my blog retweeted by a power user with 100,000 followers and another with 300,000 followers. Surprisingly, the return rate is very very low. Maybe 30-50 hits. I’ve also had some folks with 5000 followers send a RT and I’ve received 200 hits. There are lot of variables including the timing of the tweet and its subject matter, but having a lot of followers isn’t everything. Most followers are not following the tweets.

    Share
  5. Numer of followers does not matter, what matters most is who is important to you with whom you can have a meaningful conversation and dialogue.

    Share
  6. [...] but the number of times its messages are repeated–er, retweeted–according to GigaOM. So when you tweet, try to say something worth repeating. Inc.com has more ideas on how to make [...]

    Share
  7. This is so right in my blog http://socialmediadvertising.blogspot.com/ I go into depth about Social Relevancy . Its about who is following that really counts and is there exchange in conversation that matters to the community.

    Share
  8. [...] but the number of times its messages are repeated–er, retweeted–according to GigaOM. So when you tweet, try to say something worth repeating. Inc.com has more ideas on how to make [...]

    Share
  9. Klout is one. You should also look at tunkrank (tunkrank.com).

    Share
  10. Yes, Twitter followers count poor in certain cases! Here is why: When you have thousands of followers in which most of them are not real but fake/robots! I think experienced twitter marketers will analyze this section of accounts rather just updating to robots!

    Share

Comments have been disabled for this post