Klout versus Reality

Strong Arm on the Subway

“Spreading your content across multiple networks will now directly impact your Klout score. As on Twitter, our underlying premise for influence is not the number of friends or followers you have, but your ability to drive action. Create great content and engage with your network and you will see your Klout score continue to rise!”

This is the message Klout sent my friend Maddie Grant and myself when we added Facebook to our score. Great to hear successful spamming is rewarded with a strong Klout score.

Klout’s become a bit of a well-discussed defacto influence measure. But even though it takes a wide variety of mainstream social network measures — including retweeting, conversation, Facebook comments, and updates scores monthly — I wonder how effective it really is.

influence.jpg

For example, consider my Klout profile. Even though it was just updated, it lists Live Earth as one my primary influencers. This is because I RTed Live Earth every week at least once until last May. But Klout has failed to 1) update my profile to reflect current influencers and, more importantly, 2) understand why someone may be cited frequently. I was primarily RTing Live Earth because they were publishing my blog posts! So was I really influenced by myself? Even I am not that nihilistic.

There are other profiles in the influenced by/influencer of categories that I seriously question, too. I suppose that is to be expected from an algorithm. I also suppose it is to much to expect Klout to capture the zeitgeist, a moment when someone becomes suddenly influential due to an act or a situation.

There is really a lot to criticize about Klout. I agree with Mark Krynsky in particular that Klout needs data elements in addition to a better understanding of why people interact with each other. In addition, as KD Paine says, data can only take you so far, there’s a need for human analysis.

But more criticism is deserved, specifically to the communicators who have given Klout so much credence. While an interesting metric, it can tell you as much about people and their social networks as a barometer reading can tell a layperson about the weather. Consider the source.

  • Ike

    “Influenced by @laermer and @ikepigott.”

    That tells me all I need to know about your destiny as a punk degenerate.

  • I appreciate the fact that the Klout folks are working hard to make their tool better. That said, unfortunately the people relying on those scores as actual ‘influence determination’ are unlikely to make the distinction between “working hard to make it better” and “already sufficient.”

    Their metric only covers 2 sources right now: Twitter & Facebook (with LinkedIn apparently on the way.)
    But I’m commenting on a blog right now. There is no accounting for a blog? Seriously? A blog post is far more likely to influence me than a tweet.

    I am glad they are working hard – but I really wish they’d drop the hyperbole. Statements like: “Measurement of your overall online influence” only do a disservice. My online influence? Is also in comments, blogs, communities like Quora… none of which are included.

    Hyperbole. It’s the Bestest!

  • I’ve heard a lot from others that the data is out of date – and that might be an issue.

    I do have a problem with including Facebook on influence. I’m seriously thinking of weeding out a lot of my FB contacts to make it JUST friends, people that will be at my funeral. That’ll influence my “klout” but should it really matter?

    Another issue? I know someone with big “klout” that is not an influencer of many people, but tweets nonstop and RTs a lot of others. No real value to any brands.

    • To be fair to the Klout folks – you don’t *have* to have your FB score included. It’s optional.

      But yeah, the number of people you are connected to on FB seems to be inverse to the influence you have with them.

    • Jeremy, I totally agree with the concern of using Facebook for influence. It’s the one place I’ve tried to keep just family and close friends…the people as you say “would be at my funeral.” I struggle with the constant requests coming from business associates or those who do not care about my son being inducted into the National Technical Honor Society. If that affects my klout, then who cares.

  • Meg

    Suspicious about anything used to “measure influence” that relies on algorithms. Nothing is that simple OR that complicated. Influence is not a score, it’s the result of series of relationships and events and actions and reactions that evolve over time. It takes twists and turns and has ups and downs, and grows at wildly erratic rates.

    We keep trying to tell people they need more human in their brands, and then we try and take the human out of it with measures like these. I call foul.

    • I love that. Influence is not a score.

    • Love this.

      There’s a part of us which tries to quantitate everything so that we can try to predict it or recreate it. But influence? It’s qualitative, not quantitative.

    • I love this comment, Meg, but I didn’t know who you were. But your insights prompted me to click on your site to learn more. It seems like that’s the kind of “influence” people are interested in and also the kind we still haven’t figured out how to measure.

  • I think the qualitative factors of “influence” are crucial, but also the hardest to measure. We recently wrote about and posted a video that reflects those “series of relationships.. events.. actions.. and reactions”, as Meg put it. Check it: http://www.buzzshift.com/online-influencer/

  • Thanks for the mention. I was thinking of posting something similar… I hypothetically have no problem with a service that attempts to measure influence, I think that’s quite a logical objective these days. But what I do have a problem with, like you mentioned, is that the actual data on my actual profile (http://klout.com/maddiegrant) is totally out of date and/or meaningless. In my list of “most influential topics” I have… wait for it… “#FAIL” and “Google”. Now yes, I occasionally say both of those words, but I know what influences the people who follow me, I know what gets the most RT’s, and those two are not it.

    There is so much data out there – I want to see a service that actually does some real analysis of it, and I want to see one that is clear on how they arrive at that analysis.

  • I’m keeping my eye on @SpotInfluence. I think they’re going to do this influence thing really differently. Influence as defined by how likely someone is to get others to engage on a particular keyword.

  • I’ve been feeling the same way about Klout for a while. The data and influencers they have for me are out of date. In some topics I am very influential, in others, I get lost in the din, yet there is no way to determine this impact. Volume of tweets in one subject does not make influence.

    There are so many services that purport to have THE influence algorithm, but without a human critical eye, they are akin to horoscopes. Perhaps they should be used “for entertainment purposes only.”

  • Ash Shepherd

    Wonderful post and following discussion.

    I have been suspicious of Klout for awhile. Yes, I intrigued by the idea of being able quantity relationship. However, where my doubts begin to grow is that at the root of all the logic and algorithms must be built on assumptions. I don’t think the assumptions are the same between individuals on what influence plays out as in actual activity.

    Is the goal of influence to get retweeted by important people or simply get large numbers of “content consumers” to make a change int their offline life?

    When we start from different goals we can not presume to measure success or impact in the same way.

  • Really enjoyed this post and the comments.

    I agree with a lot of what was said in this discussion and have to admit the thoughts were very refreshing. Online influence is so dynamic that it can’t just be about retweets or followers or connections or friends. It has to be about all of these because if it’s not, it’s simply just an overall score of one platform. However, trying to capture every single statistic within each online platform is an extremely difficult task, but at Traackr we are tackling it one day at a time for the simple reason that this will bring the most value to the online community.

    Additionally, the idea of context and content isn’t considered in any of these scores even though this is arguably one of the most important pieces of online influence! More thoughts on measuring online influence here: http://traackr.com/blog/2010/09/no-silver-bullet-to-measure-online-influence-just-hard-work/

  • Geoff,

    Nice piece on the peculiarity of Klout. Heck, within two months, I took my Klout score from 44 to 8 to 33 and my label from thought leader to specialist to, eh hem, celebrity.

    What changed? I thought about writing a post about it but am also left with the feeling that Klout is so painfully pointless that I wasn’t sure I wanted too.

    It did make me wonder … during the month I was less active did I suddenly become uninfluential? Or was my “influential” score bogus to begin with? Give me a break.

    The bottom line is that any system that relies on frequency as a measure is flawed.

    When I called them out on this before, they gave me the same old run around that they are measuring actionable items. That is a fancy way of saying frequency, RTs, and clicks. But really, all that means is activity. Be busy on Twitter and you can easily create the illusion that you are important even when you aren’t very important at all. The contrary is true too. Important people can easily be buried by someone who RTs everybody.

    Data is fine. Giving data importance is silly. Keep up the good work.

    Best,
    Rich

  • Thanks to my friend @NotEasyToForget for posting a link to this post on my own blog.

    Some business are using (and instructing other businesses to also use) Klout for the purpose of deciding who is worthy of a response on Twitter. This sent me a little over the edge and I couldn’t let it go until I published my thoughts.

    The conversation has been relatively active, with different points of view on the need for such a tool. However, everyone agreed that regardless of a low Klout score it is unwise to ignore any user that understands how to utilize social networking tools.

    • Sounds fascinating. I read an interesting post on Beth Harte’s blog about gaming theory, which made me think of Klout. Perhaps it’s the gaming of the #s that makes it fascinating.