BrandSavant

Gaining Insight From Social Media Data

Confusing Activity With Influence

by Tom Webster on October 19, 2011

Every time I see research like this, which claims that the number of Twitter followers does NOT correspond to influence, another little piece of my soul rots away. Apparently, using “data,” the study’s authors put the lie to the “million follower fallacy” by demonstrating that users with high follower counts don’t spawn a proportionate number of retweets or mentions.

Well, I think all this proves is that social media doesn’t scale very well (and calls to mind this somewhat related piece of contorted data dredging.) But more troubling is the growing acceptance amongst marketers that activity somehow equates to influence. Influence isn’t spawning retweets. Influence is getting people to do what you want. Justin Bieber doesn’t want you to retweet him, he wants you to buy his music and pay to see his concerts. Right?

Similarly, “influence” measures are really activity measures. Else, how could they decline in the short term? It is highly doubtful that my influence – online or offline – waxes and wanes as much as my online activity does, but the most popular influence measures “punish” you for taking a few days off. In short, they treat the absence of evidence (the lack of activity) as evidence of absence – indicative of declining influence. Poppycock.

But, back to the “million follower fallacy:” I have no doubt that Twitter users with greater than one million followers spawn a disproportionately lower amount of activity than those in the middle tiers. Conversations just don’t scale like that, unless you have 20 people operating your account. It’s conversations that spark mentions and retweets, and those with greater than one million followers are necessarily broadcasters – by necessity, if not by choice.

And here is the glaring hole in the reasoning of this study that you could drive a tractor-trailer through: if the disproportionately lower activity level associated with million follower accounts somehow equates to lower “influence,” how in the hell did they get greater than a million followers in the first place?

If you can get a million followers on Twitter, you are influential by definition.

Be Sociable, Share!
Buffer
  • http://twitter.com/webby2001 Tom Webster

    And yes, I realize the “million follower fallacy” study is a year old. I can cherry-pick to prove a point with the best of ‘em :)

  • http://www.convinceandconvert.com jaybaer

    Same is true on Facebook, where many studies have shown that engagement percentage goes down as community size grows. I don’t believe it’s about influence but rather about the impact of closeness on the ability to drive behavior. When you feel like you are part of a special “club” you are more likely to engage in the behavior that the club leader suggest. When you’re cattle, that’s less true. Is it any wonder that most cults never get beyond a few hundred members…..?

  • http://www.edisonresearch.com Tom Webster

    Bang on, Jay – which is why leaders of special clubs might be more *efficient* in their wielding of influence, but I’ll never be a Scientologist. :)

  • http://www.theantisocialmedia.com/ Jay Dolan

    So, I’m not influential until I have a million followers and I make money off people sharing my stuff mindlessly?

    Clearly, I have a lot more work to do. First, change my last name to Kardashian.

  • http://www.edisonresearch.com Tom Webster

    I think that’s EXACTLY what I am suggesting that you do.

  • http://www.mltcreative.com Billy Mitchell

    It seems there is about as much debate over the word “influence” as there is about the terms “friend”, “like” and “follow”.

    As a marketer, I am unapologetic about my desire for the content I share via social media to lead to billable business engagments with new clients. Do those that follow me have a statisically higher likelihood of becoming my next dream client? Do they really, really like me? Who knows?

    As one of the top researchers on the planet with all those numbers crunching in your head, you may have better answers than me for all three questions. But here’s what I think:

    My answer to my first question is “Yes”. The odds of getting work from someone that doesn’t know me or know of me is zero. By at least being occassionally visible in their social media feed, or more likely to be found in a search, my chances increase at least to one in a million. Yes, I’m sayin’ there’s a chance!

    My answer to my second question is “No”. They don’t really, really even know me.

    My question to my third question is “If Tom Webster doesn’t know, no one does.”

    I realize that I have just had an entire discussion with myself but I invite anyone that may have been influenced to join in to please do so.

  • http://www.tommartin.typepad.com Tom Martin

    Once again, you’re nailing it Tom… it all goes back to your “do the work” mantra. Influence is just not something you’re going to be able to easily quantify (IMO) — the only way to do it is to dig in, watch and listen, and then based on the actions/activity you observe online determine who is truly influential.

    Services like Klout may say @TomMartin:disqus
    is influential on Bourbon and iPhones because I tweet about those two topics often. But what those services can’t tell you is how many folks call, email or tweet me to ask for recommendations on iPhone stuff or Bourbon rec’ds.

    That’s influence… and thus far I’ve yet to find a service that can truly measure and report that back to me at any kind of scale.
     
     @TomMartin

  • http://socialbutterflyguy.com/ DJ Waldow

    “If you can get a million followers on Twitter, you are influential by definition.” Bingo was his name-o.

  • Ashish

    This is a very knowledgeable piece of information. http://shanq.net

  • http://twitter.com/webby2001 Tom Webster

    DJ – can y0u spell that?

  • http://twitter.com/cksyme Chris Syme

    Thanks again. Whenever we need a dose of reality in the world of measurement,   you provide it. I wish there were more like you on the web–the fact that you are totally entertaining and use words like poppycock is just a bonus. We all need to understand that influence is the new “engagement”. Proof that you can skew data every which way you want, I guess.

  • http://www.meltingposts.com Aliosha Kasin

    You can spend time networking and building relationships or you can get a million followers to mindlessly share you content, Either way works I think, both take a hell lot of time and patience. 

  • Anonymous

     Tom,

    You said it, and said it well. Smart, thinking people, who may not know anything about the
    ins and outs of what we’re proposing, but they know how to make a “smart
    decision.” So  in my opinion, the job of
    influence is to persuade them over time with facts, common sense, and compassion
    that what we’re doing is something they want to be doing too.

  • http://www.thenerdynurse.com The Nerdy Nurse

    Watched your  @blogworld:disqus video on youtube.

    I have to admit, you are very charming and witty!

    You have me even more excited about @blogworldexpo 
    kudos to you for that!

    What do you recommend as the best method to put a brain fire out?

  • http://twitter.com/andreaswpv andreaswpv

    Like your tone of voice, but not sure I would give in that easily. 
    1. Just because you got to 1 million does not prove anything – you might just have changed your communication when getting more and more connections
    2. Relevancy has something to do with impact and by that with relevancy and content.  While I totally agree that we usually measure activity and not influence, it still looks like your conclusion that number of followers indicates impact is missing these elements.  

  • http://twitter.com/Collectual Collective Intellect

    I don’t think it’s the concept of tying online activity to influence that’s misleading but trying to use a single measure of activity to indicate influence. I don’t believe an organization should rely on a single metric, out of context, to define influence and ultimately influence how they engage with their customers. There are some may characteristics that define a user’s online behavior that understanding what you are measuring (i.e. Viewing Intent, Customer Service Sentiment, Demographic/Geo data) will ultimately categorize consumers based on those analytics guidelines. Once this audience segmentation is defined, a set of influencers may emerge but that group may shift as the context does.

  • Sean R. Nicholson

    Influence? What did Charlie Sheen do to be influential to gain hundreds of thousands of followers? What did the Bronx Zoo Cobra do to become so “influential”. Don’t confuse popularity or ability to be humorous with influence. Oprah is influetial. She recommends something, peoplebuy it.

    If Charlie Sheen recommened something, I’d probably run the other way.

    Just because you have the ability to stand on a soapbox and shout loudly at the wind, doesn’t mean you influence anyone. And there are plenty of spam Twitter accounts out there with 25,000 spam followers to show that you can buy a large number of followers, these days. Just because they have a large following doesn’t make them influential.

    Think about it. If you follow someone that has 50,000 followers does that mean they automatically influence you? If so, you’re a Twitter sheep.

    A number is a number. Just that. And Klout and Flout just track people’s abilities to stand on a soapbox.

    My $.02…

    –Sean

Previous post:

Next post: