Pages

KLOUT doesn't suck

Back in the days before computers, the question of what something was worth involved a lot of intuition, instinct, feelings in the gut. Prices for everything from real estate to stocks to bread went on the basis of what people kinda thought was about right, based on the costs of inputs and guesses about what the market would bear.

Then came the quants. A quantitative analyst is a person who works in finance using numerical or quantitative techniques. No longer did you have to wonder or guess or estimate how much something was worth. You could calculate it. Identify the inputs, quantify their costs under relevant conditions, model their respective sensitivities (i.e. how much each contributing factor actually influenced the value), and - PRESTO! - you know how much something is worth. Lots of things for which it used to be impossible to assign a solid, quantifiable valuation were suddenly laid bare to the market.

How much is rain worth? What does the flu cost? What benefit is there to making people happy?

All of these yield to quantifiable analysis. When the quants get it right, things are valued properly and efficiency goes up - fortunes get made. When they get it wrong, things are overvalued or undervalued and efficiency goes down - fortunes get lost. As quantitative analysis got more sophisticated, it started looking at valuations for consumer behaviors, not just vendor goods and services. 

Enter Klout.

Klout is a system of measuring and valuing influence online as result of interactions on social networks, such as Twitter, Facebook, LinkedIn, etc. There have been plenty of other metrics for measuring this, or attempting to. Some, like Twitalyzer, persist and compete with Klout; others, like Twinfluence, dried up and went away. Klout has gotten a lot of heat lately, for their corporate policies and for their adherence (or lack thereof) to various laws. These are legitimate criticisms.

For some, however, Klout bears a painfully close resemblance to the queen bee arbiters of cool in high school. Klout is, by this thinking, run by capricious, self-appointed, evil douchebags using some secret formula to assign a number to something that every right thinking individual simply KNOWS cannot be quantified. After all, anyone could just make up some competing arbitrary metric a priori and apply it. It would be just as valid, no?

Well, no, actually, it wouldn't.

Meet person A and person B, Alan and Bob. Alan and Bob both spend the same amount of time on the same social network platforms. Both have similar numbers of connections, both do roughly the same amount of talking (or tweeting, posting, flapping, or whatever). Each of them has a social network, an online presence where he is known.

Now then, the big question: whose social network is more valuable? One talks mostly to his friends and family, the other talks mostly to the leaders in his chosen fields of personal and professional interest. One has a lot of acquaintances, the other has a few close friends. One retweets lots of interesting and informative links, the other is funny and original. Whose network is more valuable? And valuable for what purpose? In what context?

There's an old adage in advertising. It says that half of the money people spend on advertising is wasted, but nobody knows which half. Wouldn't it be nice to know which part of your time, money and effort was being spent effectively and which was being thrown away? I'm not talking about Walmart or Starbucks or Toyota, here. I'm talking about anybody who has something they want others to buy, such as an anthology of well-written flash fiction. In a focused effort at finding buyers and clients, if there were some way to know which people are a dead end and which would lead to repeated sales, wouldn't that be valuable information? If I'm trying to sell fiction, Merlot or movie tickets, wouldn't it be nice to know who are the opinion leaders in writing, wine and entertainment? Klout isn't something that just OTHER PEOPLE can use, you know. You could use it too. If you wanted to.

Klout is certainly secretive and self-appointed. Klout is certainly ranking you and selling your contact info to people who want you to buy things. Klout certainly decided what to measure and how to value each component, and they probably made mistakes in their quant model. Klout may also be evil, in much the same way that Facebook and Google are evil.

But is Klout pointless? Capricious? Arbitrary? No. That's a simplistic misunderstanding of what quants do, and how Klout is using quantitative analysis to figure out what your social network is worth.

===== Feel free to comment on this or any other post.

9 comments:

  1. A very, very interesting post. I read it carefully. Paused a bit and gave it some thought. I wasn't unfamiliar with the basics of what you wrote - what got me thinking was how you compared Bob and Alan. As a former high school teacher, the same process was used to determine whom to nominate for various school-related activities. It didn't always work, i.e., we'd think afterward, perhaps we goofed and Bob would won the contest. But we had to have a baseline, where the tangible and intangible meet.

    ReplyDelete
  2. @ Kittie: I'm glad you found it thought provoking. Things like this are very difficult to quantify, which is why some people mistakenly think they are impossible to quantify. The valuations for each part of the model are decisions made by people, driven by particular goals and agendas. Mistakes and biases are part of the minefield, as is the second-guessing after the fact.

    ReplyDelete
  3. I keep seeing lots of people making $$$ off Twitter, but Twitter isn't one of the major benefactors. The thing is, if QA is all that and a jar of peanut butter, why are people always publicly questioning their high scores for topics they almost never tweet about?

    The other thing is, Klout is scraping the networks and making money by putting ad targets on people, but what do we get in return? A number?

    ReplyDelete
  4. @ FAR: why are people always publicly questioning their high scores for topics they almost never tweet about? I think because it's funny to do so. Klout thinks I'm knowledgeable about vaccines because of something I RTed? That's wacky, so I tweet about it. However, Klout also thinks I'm knowledgeable about fiction, blogging, writing and humor. Those are accurate, so I don't make a big deal about them. As quick check of Klout tells me that you are influential about writing, authors, science fiction and Georgia. Accurate? Or not?

    What do we get in return? Excellent question. We get a clearer picture of what impact our activities online are having. Granted, for you, me and perhaps for most people, the nuts and bolts of this is of little to no interest. The number is little more than a line scratched in the dirt for a backyard pissing contest. However, there are people, entire industries in fact, which live and die by this kind of information. For them, a clear sense of what is connecting and having a solid impact is terribly important.

    Klout is measuring you whether you want them to or not. Controlling how they do it means being part of the process.

    ReplyDelete
  5. I have to confess that i'm kind of a fan of klout. Not in a "heh heh, i'm gonna rule the world" kind of way, but in a number-crunching quantitative goodness kind of way. Social media is such a vast and subjective area that the very idea of trying to analyze it using an algorithm is fascinating for a math-geek like me. Much of it is highly suspicious, like when i tweeted about some kids who had set up a lemonade stand near my house and i got a few rt's, suddenly Klout thought i was influential in lemonade. There are holes that you could drive the Titanic through in the algorithm, some are demonstrated in this article: http://techieminx.wordpress.com/2011/11/13/this-makes-no-sense-comparing-klout-scores/ And although i don't like the privacy angle (or lack thereof) one little bit i've gotten a couple of little perks (a free ebook, a discount code) i figure my tweets are public anyway and with the lack of privacy on facebook and g+ i treat those as public forums anyway, so no harm done. Interesting post, Tony - take Klout with a ginormous grain of salt and it's interesting to check out.

    ReplyDelete
  6. Thanks for the post - It reminded me of a tweet I made on Nov. 11.

    I check #FF carefully - Twit. should tabulate click-throughs & other data 2 encourage quality instead of quantity. #shotgun tweets :-(

    (I did write a much longer comment, but it was rambling... lot of thoughts on the subject of quality vs quantity.)

    ReplyDelete
  7. I would have liked you to introduce Klout sooner in the essay and likewise mounted your defense of it sooner. As it was, I was only ready to hear you out when you closed it. It does sound very much like the obnoxious Queen Bees of high school, except with more monetary implications, and no necessary merit to that. With my experience of Klout, and from what you've said about it here, I see no tangible or exemplary ways in which their analysis of anything is better than what other companies have been using for years. I'm sure you have such details to support your thoughts, though, which is why I wanted to read about them. Especially since we're approaching this quantification (which in many ways I find disgusting) in quantifiable ways, a defense of Klout really requires data. Do frequent users make more money? How successful are top-scoring Klout clients? Do most people who rise in Klout-score see real-world increases in success? Is data in Klout more accurate than a person's referral numbers, or credited sales, or hits and follower count? And it must go beyond anecdote - any one person in a crowd can succeed - but have statistics to back up that Klout actually causes significant success.

    Having said all that, I'll suppose my score is very low.

    ReplyDelete
  8. @ PJ: The number-crunching aspect is what intrigues me, as well. This kind of valuation is still difficult and highly dependent on the assumptions of the modeler. Social networks have been around for 15,000 years, but only in the last decade or so has there been any way to get lots of data about them.

    @ Bree: Tabulating the click-throughs and other behaviors is something Google Analytics does, but only to the extent that it can see activity on different websites. That's why Klout encourages you to integrate all your social networks, so it can do just what you're proposing.

    @ John: Thanks for reading through the tedious part at the beginning! Actually, if I had that kind of rigorous statistical analysis of a large, objectively obtained dataset, I wouldn't have written a blog post - I would have written a freelance article for Forbes. Is Klout succeeding, i.e. creating a meaningful metric? As I noted, they are neither the first nor the only company to try to do this kind of analysis. Are they any better at it? Do high Klout scores correlate with sales and profitability? I only have anecdotal data, both pro and con, neither of which would be enough to dispel your disgust.

    Having said all that, I'll suppose my score is very low. No, not really. You've ignored Klout completely and will presumably continue to do so. Has your Klout score suffered for this? The mistake some people make is to focus on the score itself instead of the social activity. Doing all the little tricks and shortcuts which bump up a Klout score, or a Twitalyzer score or whatever other score, is just gaming the system. One can use the numbers to get a better understanding of behavior, and that's all their good for. Basing self-esteem on them, ala the Queen Bee analogy, is a mistake.

    ReplyDelete
  9. Excellent post, Tony. Although I'm appreciative of the +Ks that have come my way, the whole concept makes me feel uncomfortable.

    ReplyDelete

Thank you for leaving a comment. The staff at Landless will treat it with the same care that we would bestow on a newly hatched chick. By the way, no pressure or anything, but have you ever considered subscribing to Landless via RSS?