The Trouble with Klout

Estimating influence is a delicate balance of art and science. People are drawn to quantitative methods because scores are easy to understand. The downside of reducing influence to a number, though, is oversimplification.

Paul Gillin's Klout InfluenceLately, I’ve been looking at Klout, the popular new tool that bills itself as “The Standard” for influence measurement. The more I look at it, the less I like it. Klout’s weaknesses have not stopped it from amassing an impressive list of more than 3,000 business customers and from being incorporated into popular applications like HootSuite as a standard metric. It is “the emerging standard” for measuring influence online, said Klout Marketing Manager Megan Berry in a podcast interview with Eric Schwartzman last month. I just hope those clients aren’t taking this metric too seriously.

Beyond Followers

Klout attempts to determine influence metrics by looking at a person’s online activities and the actions of others that result from them. The thinking is that influence isn’t a matter of how much you say as much as the impact your words have on others.

Many people have a Klout index and don’t know it. The service crawls Twitter and ranks members automatically. If you want to grow your score, you can log in to the site and give it a bunch of information about your online activities. I spent 15 minutes on Klout registering my social networks and grew my score 10 points on the spot. This is a major flaw in Klout, but more on that later.

Klout uses a proprietary algorithm to estimate influence based upon comments, retweets, @replies and mentions, among other things. The company isn’t very transparent about how it calculates the score, and with good reason. The algorithm is a competitive asset and disclosure would inevitably invite people to manipulate the system.

The downside of opacity is confusion. By revealing so little about how its ratings are calculated, Klout essentially asks customers to put their faith in the service to do the right thing. This is dangerous, given Klout’s flaws. Nevertheless, the score is a public record that anyone can see, and its influence is growing to the point that Klout scores are now reportedly showing up on resumes.

The Shirky Effect

Clay ShirkyThe problem is that some of the ratings are nonsense. For example, my Klout score (66) is modestly higher than Clay Shirky‘s (60) and significantly higher than Marc Andreessen‘s (42). This is ludicrous. Shirky (right) is the author of two influential books about online sociology and has been a thought leader on the Internet since the mid-90s. Andreessen (below left) invented the browser, cofounded Netscape and is one of the fathers of the modern Internet. Both are sought-after speakers and the subject of extensive Wikipedia articles. Yet Klout says I have more influence.

Marc AndreessenThe problem is that neither of these brilliant innovators plays by Klout’s rules. They aren’t active on Twitter and they don’t have Klout accounts. The fact that a single post on Shirky’s blog can draw more than 1,200 comments or that Andreessen’s occasional writings appear in The Wall Street Journal is of no consequence. Klout doesn’t monitor either of those outlets.

Klout’s bigger flaw is that its scoring system is tied to membership. The more you tell Klout about you, the higher your score is likely to be. This linkage fundamentally undermines the quality of the service. In effect, Klout pays you to endorse its service by rewarding you with a higher rank. If Google did that, Congress would be holding hearings.

A Million and One Improvements

Klout admits that its methodology isn’t perfect. In the interview with Schwartzman, who is the co-author of my B2B social media marketing book, Megan Berry said the company has “a million and one” improvements it wants to make. Schwartzman pressed Berry hard on shortcomings in the Klout methodology, and her responses were a weak defense. In essence, Klout treats every social network the same and all interactions equally, she said. A retweet, which is a one-button operation, is just as good as a thoughtful commentary on a blog. Except that Klout doesn’t currently monitor blogs, other than those on Google’s Blogger service. That must be one of the million-and-one improvements in the pipeline.

Megan Berry on KloutA comparison of Berry’s and Schartzman’s Klout profiles showcases the service’s flaws.Berry’s Klout score as of this writing is 70, while Schwartzman’s is 60. Barry does have a couple of thousand more Twitter followers than Schwartzman, but she said Klout ignores follower metrics as meaningless. Berry is very active online, but not nearly as active as Schwartzman.  Her blog has been updated eight times this year while Schwartzman has posted 36 episodes of his popular On the Record…Online podcast and more than 30 entries on his Spinfluencer blog. Berry contributes occasionally to Huffington Post and Mashable, but Schwartzman is also active outside his own channels, contributing to Social Media Today and For Immediate Release. Schwartzman has 44 recommendations on LinkedIn, while Berry has three.

Eric Schwartzman on KloutAs far as I can tell, there are two principal reasons why Berry outscores Schwartzman on Klout. One is that she knows the system. She has at least a vestigial account on every social network that Klout cares about, whereas Schwartzman limits his activities to fewer outlets. Berry also tweets regularly on behalf of her employer, giving her Twitter account a Klout halo effect that attracts retweets and @replies.

My intention isn’t to pick on Megan Berry. She’s obviously a bright young woman who’s very savvy about social media. However, there’s nothing I can find that qualifies her as significantly more influential than the veteran Schwartzman, not to mention Marc Andreessen.

In her interview with Schwartzman, Berry described Klout as “[Google] PageRank for people.” In my opinion, it’s got a long way to go. Klout has some utility as a way to compare the online presence of active social media users, but measuring influence is much more complicated than counting retweets and Foursquare tips. Klout is betting that it can use its metrics to entice (coerce?) people to join its social network, which it can then monetize through advertising. The link between membership and Klout score is a disturbing weakness. Proceed with caution.

38 thoughts on “The Trouble with Klout

  1. substitute the google plus suggested user list here. add to it the inflated protesting of those who proclaim they don’t want to be on it. etc. nonetheless, were I in the biz of hawking myself as a sector authority (which neither shirky nor andressen really having to worry about), I would want to play that game. most of this stuff is for the fishbowl experts, putative legends within a narrow population. ok, time for more coffee and less cranky.

  2. An excellent post and analysis but you are missing the point here. Klout, like all social media, is about achievement in the online space. There is no real connection between offline and online (unless we somehow manage to beam the constant data feed of the Internet into people’s minds while they sleep). It’s a game. Klout itself is a “badge” that people can earn. And they can see it increase by “playing the game” (according to the rules set by Klout). Will there come a time when the online world and real world are blurred enough to make a service like Klout apply to both? Perhaps. But it won’t happen for quite some time. Until then, people have to actively choose to participate in the online world to increase their “engagement” and “influence.” Remember that I can’t be influenced by Andreesen if there is no way that I can connect to him (i.e., I don’t read his books, I don’t attend his conferences, I don’t hear from him via Twitter, or email). It’s a relationship that is created between he and I online which Klout measures.

  3. Great points Paul.
    Basically Klout measures tweets, that’s it. They use follower count to calculate their True Reach, by the way: http://www.martijnlinssen.com/2011/09/klouts-true-reach-simply-your-follower.html

    Klout however keeps spreading the nonsense that they measure online influence. They don’t. They market the idea that they do, while simply counting tweets. Retweets and @replies determine your amplification, unique people who retweet and @mention you determine your network

  4. Thanks, Jason. I see your point but Kaloutsky doesn’t position itself as a game. It refers to itself as the standard for influence measurement. It doesn’t even qualify that by saying “online” or “social media” influence. It has set expectations very high for the quality of its rankings, but its current methodology is flawed. My issue is not with its tactics but with its lofty definition of its own value. Thanks for commenting.

  5. This is a sound critique of Klout as “the standard of influence” and I agree with it, generally speaking.

    But what if Klout measures “klout” – a new kind of influence – and not, in fact, traditional “clout”.

    Influence is constantly being redefined and the standards for measuring it are always evolving too.

    I think Klout, with all of its limitations, is in the process of remaking our understanding of influence. For better or for worse.

  6. I tweet for a number of work related accounts via Hootsuite. I find Klout hilarious. On one account, I’ll gain 25 new legitimate users in a day and have my tweets retweeted 5 or 6 times and my Klout goes down.

    I’ll have no new users on one account, have nothing retweeted and have tweeted once in 3 days and my Klout will go up.

    I started a new account for a new product, had 15 followers and had a Klout of 26 right away – how is that even possible? I had no clout with that account at all.

    I ALWAYS want to dis Klout on Twitter but don’t want to upset them :-) They say they may not be counting followers and I call bullshit – sometimes if I get a big hooker surge, I get a higher Klout score.

  7. Paul, I think the biggest flaw in this post is going after the Klout score of a single person, particularly one who is connected to the company itself, as an illustration as to the flaws in the system. There are much broader issues – even among people who are active in social media – that illustrate the much broader weaknesses in the system.

    But the biggest question I would have about Klout for me is, “Why would I care?” As an end user, do I value tweets more from people with higher Klout scores than I do those with low Klout scores? No. I value tweets more from people whose clout in MY mind is stronger, and there is unlikely to be a metric for that anytime soon. For marketers, if you are out there thinking, “Gosh, if I just target the people with the highest Klout scores I’ll really knock it out of the park,” then, well, don’t put too many photos on your desk – because it will take that much longer to pack it up when you get let go.

    Klout, Empire Avenue, and these other games of “measuring influence” are just that – they are games. Their greatest value lies in that they create conversations about what “influence” really is, and that is actually worthwhile. But beyond that, it’s a game, and I would hope savvy people in the space take it for what it is.

    But Gillin, c’mon man! You’re just one point behind @ev and two behind @biz – and they co-founded Twitter, so what do they know? Now if you want to compare their 67 and 68 to Ms. Berry’s 70…

  8. LOL! Good points, Ted. I should point out, though, that I didn’t go after the Klout score of a single person. I included Clay Shirky, Marc Andreessen, Eric Schwartzman and myself. In all cases, I believe out Klout scores are out of whack compared to out true influence. A couple of other commenters have made the point that this is a game, and that’s fine. Just label is as such. When you call yourself a “standard,” though, you raise your profile to a higher level. If Klout were a bit more humble in the way it portrays itself, I wouldn’t feel compelled to point out these flaws, but the company’s own marketing demands that it be accountable to its claims.

  9. After tweeting for all of three (3) weeks about a narrow section of the travel industry, I retweeted – exactly ONE time – about a former colleague’s request to vote for his proposed workshop at SXSW. Then I checked my Klout score. It seems I am influential about SXSW and nothing about travel. Can I say … BS?

  10. Your last comment hit the nail on the head. It seems to be Klout’s lack of humility that attracts this criticism upon them and their algorithms. I will add that it also seems odd how they identify the topics upon which an individual holds Klout. For example, my (@BLichtenwalner) lowly klout of 44 is apparently all targeted at topics of “Branding”. I can’t remember the last time I mentioned branding…

    That said, to their credit, Klout seems to be the forefront in this arena – even if it is just a game. One big positive step forward is their (Beta) ability to assign +K to individuals you choose. Although it is buried: Search for user -> Topics -> Click “Give +K”. It’s also only available for topics they choose. I guess I’m just branded as the branding guy for eternity…

    Now, if Google shifts their +1 to allow you to +1 people, maybe we’ll know what their Glout (too close to “Gout”?!) is – a score with some real clout behind it.

    Great insight – thanks for sharing Paul. By the way, I just gave you a bunch of +K – particularly around Klout. You’ve earned it.

  11. Barbara: Thanks for a story that reinforces my major point. If you don’t know how the system works, how can you trust it?

  12. Paul,

    We’re totally in the same camp that there are flaws within the Klout approach, notwithstanding the fact they have enjoyed tremendous success. At Appinions, we take a different and unique approach to identifying influencers by focusing on topics rather than applying scores based on activity. Our Influencer Exchange also takes into account not only social media but traditional news sources, and radio/TV transcripts to provide a comprehensive view of the influence landscape.

    Larry Levy
    CEO, Appinions
    Appinions.com

  13. Paul,
    A very good post – thank you. I took a hard look at Klout when it first started and then again about a year ago and came away both times feeling that it didn’t live up to the promise. You and others have pointed out many of the flaws but I have always thought that it missed the relative *insignificance* of a million people retweeting Justin Bieber because he is popular in culture in general v. retweets in smaller numbers that are modified (showing more active user interest/thought). I completely agree with the impact of bloggers and the comments they entice being missed by Klout – this is a big miss. But, I think Ted said it best – “why should we care?”. We are getting close to a day when content will come to us because of a combination of our interests and text analytics’ ability to parse content for meaning and align those two. Once we get to this point, Klout, like many others, will just be a “twitter popularity meter” (which is what I see it as today). Klout is not bad (IMO) but it is selling itself as more than what it really is.

  14. Klout’s origins are actually as a Twitter popularity metric, and in that respect its methodology isn’t bad. The problem is that it now proclaims itself to be a broad measure of influence yet doesn’t index mainstream media or WordPress, not to mention LinkedIn and an assortment of other social networks. I trust those enhancements are coming, but until then the service should tone down its claims.

  15. Excellent post that raises questions about the growing importance of influence and the lack of one particular tool’s ability to accurately measure the same. I also find Klout lacking in clout. Among the most obvious shortcomings re my score (57); Klout constantly identifies people as having influence over me, yet I don’t even know who some of those people are. What Klout does have is an excellent business development and marketing operation. 3000 business partners isn’t too shabby, but the company had better improve its algorithm because their credibility is on the line. However, as long as people continue to misunderstand what it is that Klout is actually measuring, a slice of someone’s life online and NOT the totality of who they are and the influence they actually yield (see Clay Shirky & Marc Andreessen), the company is going to continue to gain “clout.”

  16. I remember the flurry of commentary that accompanied the release of Klout and its hyper-promotion by Mashable. It was a mostly a congratulatory rush to get with the latest new thing. Anyone who was anyone ‘social’ was registering to accept their “influence” score and trade on a new social currency before someone else did. After signing up and exploring Klout myself I remember thinking:

    1. Oh, crikey! Another Hubspot-like marketing ruse offering a free “grading” for something. Except they wanted to grade me!
    2. Nice website design and UIX.
    3. Damn, I wish I had thought of that.

    After the dust settled (in about two weeks), I noticed Klout was just another player in a bigger trend of Personal Vanity Metrics (PVM). PVMs are so popular because they gamify the online experience with a competitive Leaderboard that lets people aspire to win and feel special. Others include Smarterer and Work-for-Pie.

    I agree with Paul Gillin that they are mostly irrelevant self-congratulating metrics. But, as Jay points out above — they are probably new metrics that needs to be given a place at the table. Not just for their obvious popularity with the socialrati, but because advertisers ARE listening too. =P

  17. So, what about building a better mousetrap?

    I’ll assume that you believe measuring social influence has value.

    What things do you think should be measured, how do you think they should be weighted, how do you think they should be automated. I say automated, because doing this with human intervention for a website which gives the information away for free is probably not a profit-generating idea, at least not initially.

    If Klout doesn’t have it right, in other words, what IS right? No doubt there is much disagreement on this, but this blog makes great, valid points. How can a system be put together in a supportable, sustainable, revenue-generating way that addresses these issues?

  18. And may I say? Although what you’re saying here is extremely accurate, I’m a bit disappointed to realize it isn’t worth anything on the day I found out my Klout score was 72. Just sayin’. :)

  19. A very interesting piece. While I agree that Klout scores are not the end-all, be-all when it comes to measuring influence, I have to agree with what a lot of people are saying here. It’s a game to measure your social media influence, that’s it. It shouldn’t be taken too seriously and unless you’re applying for a job that requires a social media influence, I don’t know why someone would want to put it on a resume! Personally, I just use mine to track how active I’ve been in social media and when it dips, I take it as a sign to step up my game a bit. Great article though, it definitely sparked some interesting conversation!

  20. Thanks, CJ. I’m not sure influence CAN be accurately measured without the human component. One of the principal weaknesses of services like Klout is that they can’t automatically federate your online identities. My other blog, Newspaper Death Watch gets far more traffic than this blog, but could an automated service figure out that that blogger is me? I actually think Klout is right in offering people a chance to fill out their own profile and tell the service where else to look for influence. Where it falls down is in pretending that that step isn’t necessary. I honestly don’t see how a truly reliable metric can be created that doesn’t involve some level of user-intervention. I have no problem with that as long as the methodology is transparent. My problem with Klout is that they claim to be something they aren’t.

  21. Thanks, Alexandra. My problem isn’t with the Klout methodology but with the implied authoritativeness with which the metric is presented. If you look at the customer list, you see some very heavy hitters have signed up to use it. Are these companies aware that this is a game that shouldn’t be taken too seriously? As I’ve been writing here, my issue is with the way Klout presents itself more than with its number.

  22. Good post, Paul. I’ve felt that Klout is interesting but flawed. According to Klout, I’m influential about topics I don’t generally tweet about (including photography, parties and Friendster). So if they get that wrong, it’s hard to have confidence that they’re getting other things right.

    Also, as I point out in a blog article, http://bit.ly/mS3FMc, it’s possible to boost your Klout score just by tweeting a lot. Earlier this year, my name was mentioned in some spammer’s tweet, and when I checked out the person, I found that he had only 11 followers but a Klout score of 59. Why? Because he had posted 9,156 tweets.

    Seems it’s very easy to game the system to get a strong Klout score by posting a lot of tweets — regardless of whether anyone ever retweets or responds to your tweets.

    In terms of influence metrics, I find TwitterGrader to be a more interesting tool than Klout. But we still need a better, more comprehensive way to measure the impact of social media.

  23. I have recently used Klout for a company I work with in a promotion and I have to say that I loved the results. We targeted a certain amount of people and they engaged with the client’s product as well as shared the client’s product with their user base. While I am under no illusion that I reached the definitive top influencers in this category by using Klout I am really happy with the results.

    While I don’t think they have absolute authority, I think Klout gives enough of a quality sample set to work off of for a business. (for a business I think this is a selling point) However, if I was a company looking for the top five influencers in a niche I wouldn’t judge that by Klout score.

    In the end if you want to take a shotgun approach to amplify something to a group of people that are probably at least moderately influential in a specific niche Klout did the job. If you want to use a scalpel to find your influencers I don’t know if it is the metric you want to use.

  24. Awesome article! I’ve recently checked out Klout and still can’t figure it out, not the big deal with it. Are we so easily influenced or do we just need to “feel” important?

  25. Paul, you’ve made a number of great points here. Enjoyed reading this and lesson to all who may tout their klout. It’s also misleading to think folks like Michael Jackson have a score of 75 (he’s not alive by the way — not sure when he last blogged, tweeted, or engaged in conversation).
    In any case, the perspective I recommend others using for klout is that it’s 1 measure of a person’s social footprint. I use it frequently to assess personnel, new business partners, firms, etc., but I also look at other details like – LinkedIn profiles, Google +, FB, Twitter followers & # of times listed. This gives me a perspective on how they are engaging, on what topics they may be influential, and their affinity for social marketing.

  26. Paul
    Excellent Post! I really like how you articulate your points. I really like the point of being a member and talking about yourself gives you more points. It is a way for Klout to use its members to tell others about its services and maximize its profit potential through other services.

    The only thing I know is that the result of my Klout score going up or down has no impact on my website analytics. Till that happens, I can’t acknowledge Klout as the standard for influence. I will take quality tweets, posts and content over quantity tweets, followers, posts and content any day.

  27. Jason writes “I can’t be influenced by Andreesen if there is no way that I can connect to him (i.e., I don’t read his books, I don’t attend his conferences, I don’t hear from him via Twitter, or email).”

    And Jason doesn’t use a browser or the web…

    Tim Berners-Lee (@timberners_lee) Klout score is 47. My score is 48. It’s ludicrous that any of us should rank higher than the man who invented the web.

    Klout is a good idea. But it has serious flaws.

  28. My Klout increased dramatically because 3 individuals I know started following me, they have a following of some 200 000 each. My score then proceeded to leapfrog most established online stores, software companies and media sources. Klout, is essentially, a joke. I have very few followers on twitter, some 200, but because those followers have incredibly high numbers makes me influential? Even though they do not retweet, reply or mention me? Makes no sense. My friends on facebook are influential people in their industry, that doesn’t automatically make me influential, Klout.

  29. That’s a great example of the limitations of algorithmic approaches to measuring influence. I don’t quibble with the details of how Klout measures influence. My problem is with their claims that they are some kind of a “standard” for that measurement.

  30. Paul, I think your initial comment about oversimplification is the key to this discussion. Klout is a tool, just like the multitude of other tools out there. If you know what it is, how to use it and what its limitations are, then it does provide value in the toolbox of influencer analytics. When the complex definition of “influence” is boiled down to a single, definitive number, that’s when you run into trouble. In my experience with Klout, it works as it is designed: it does a decent job of evaluating current influence across the social ecosystems users have told it they belong to. In my evaluation, Klout score is not tied solely to number of followers and I do notice that when my number of personal interactions with other influencers subsides, my Klout score starts to go down, just as would be expected. That said, as you note in the comment above, it takes human judgement to put the complexities of the Klout score and its components into context with other analytic tools to draw a complete conclusion on who are true influencers and who are opportunists. No single score can do that.

  31. Great post with excellent comments.Some Klout scores seem to be dropping all over the net lately (5th Oct),many people asking how come? It seems Klout may have some new bed fellows (Google+),and with paid advertising on Twitter starting next year I think the ‘Standard of Influence’ may be the new “Standard of Effluence”

  32. I agree with you that Klout does a pretty good job of assessing influence within its narrow ecosystem. The problem I have is that its slogan is “The Standard for Influence.” It’s a standard according to whom? Ho standards body has bestowed this honor upon it, and its shortcomings are pretty severe. I don’t have a problem with what Klout does as much as with how Klout positions itself.

  33. Pingback: Mide tu influencia online y dime de qué pasta estás hecho « Blog Every Week (#blogeveryweek)

  34. Pingback: Sourcing3 Buyer & Supplier Magazine - Best of B2B Marketing Zone for 2011

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>