Five Questions Klout Can't Answer

rohnjaymiller
Rohn Jay Miller Director of Digital Strategy, Hanley Wood Marketing

Posted on November 29th 2011

Before moving on to more pressing issues in social media—like Google+, Facebook and the emerging Social Enterprise, to name a few—let me nail to the wall this one final post about the disaster called Klout.

After my initial post “Delete Your Klout Profile Now,” last week I offered up a dozen questions and an open mike to Klout CEO Joe Fernandez (“Klout CEO Responds to Critics”) which he graciously stepped up to.

Except, well, Joe didn’t really answer the big questions that were asked of him by myself and many other bloggers after revelations that Klout was tracking friends from Facebook with private profiles, sweeping up minors and publishing Klout scores on them, and generally prevaricating about the real purpose of Klout and how it deals with the clients who pay the bills—advertisers like Chevy and Virgin.

Here are five questions that remain unanswered by Fernandez and Klout.  And I don’t expect clear, accountable answers on any of them any time soon:

1. What is the precise number of profiles Klout has in its systems of people who have never registered with Klout or opt-ed in to Klout in any way?  And are you deleting  these “unrequested” profiles from your database?

Klout sweeps the Internet for public social media messages—right now that’s from Facebook and Twitter.  (They also will allow people who register to add links to their LinkedIn and Google+ accounts.)  Klout claimed in September to have more than 100 million profiles.  How many of these are people who never registered at Klout, and never asked for a Klout score? 

This is an easy question to answer (subtract total profiles from total registered profiles) but not easy to make public.  If Klout admits that 90 million of the 100 million accounts are people who never asked to be rated by Klout, privacy issues here and in the European Union become a very public issue.  Klout doesn’t want to talk about who’s in their database and how they got there.

Think of Klout as “anti-social media.”  It’s not about sharing and building organic networks of relationships using social media.  It’s turning social media into a popularity contest, claiming a Google-like right to mysteriously—and publicly--- rank your “influence.”  They might succeed at this if they go unchallenged as they amass information and use it for a proprietary public ranking.  They’re aiming to define and own a space called “public social influence ranking.”  But their lack of transparency and  accountability in publishing public rankings of people who largely haven’t opted-in will be their un-doing.

2. How many minors under the age of 18 have Klout profiles?  How many under the age of 13?

Klout can’t answer this one because there’s no way to automatically identify these profiles.  There are still underage kids getting Klout scores and being identified as experts with “social influence.” 

Fernandez’ answer was mush mouthed political candidate-speak:

“This is a challenge that every company doing business on the social web faces. We work closely with the platforms and their trust and security teams to share insights and best practices. I think there is also a role we can play here in helping parents understand how data is spread on the social web so they can be more informed about what their kids are doing.”

In other words, “we don’t know, but  that’s really a problem for parents, not us.”

3. How can anyone trust Klout scores when you keep your algorithms secret and you won’t allow an independent third party to audit their accuracy?  How can you claim transparency when you’re not transparent? 

These people want to be “the standard for influence.”  Everyone from the Gallup Poll to Neilsen to the American Bureau of Circulation provides clear, unambiguous information about their processes for polling and reporting and the raw data used to compute the results.   There’s a dozen ways to game the system and there’s no discussion about how Klout proposes to police the “black hat” and “grey hat” tools and techniques to goose your Klout score.

4. What’s the specific number of people who have opted out and deleted their Klout profiles? 

Joe won’t answer this other than to say “it’s less than .01% of all profiles.” So there’s no hard numbers.  But note how he answered the question.  Klout has developed profiles on 100 million people, mostly using public data and without their knowledge and opt-in permission. 

As small as .01% is, out of 100 million profiles that still would represent 10,000 people who have deleted their profiles since November 1st, and I think it would be a reasonable guess that most of those came after the fusillade of articles and blog posts in the middle of the month. 

Klout could be losing 10,000+ profiles a month---and among the most active people in social media.  A study by Cornell University and Yahoo in March 2011 pointed out that 50% of all tweets on Twitter are sent by just 20,000 people.  Think about those numbers for a moment.  Even if the Cornell-Yahoo study is wrong by a factor of 10, that still means a few hundred thousand people make up the majority of tweets on Twitter.   How many of these “super users” are among the 100,000 or so who have deleted their profiles?  That specific question goes unanswered.

5. Why is Klout hiding the “delete your profile” option so that it can’t be found?  

There are only three hidden ways to delete your profile on Klout.  The easiest is to follow this path:

  1. Log in.  If you don’t have an opt-in account, you’ll have to create one using your Twitter or Facebook profile.
  2. Go to”profile settings.”
  3. Scroll to the bottom of the page, which is hidden below the fold.  You’ll see these words: “Klout values your privacy. Click here to learn more.”  Click on that link.
  4. You’ll arrive at the “Privacy Policy,” which is three screens of boilerplate privacy verbiage 1,259 words long.  At the very, very, very end of the Privacy Policy it says: “If you are not a Klout user and wish to opt out of Klout, please click here.”
  5. You’ll finally arrive at a three page dialogue to delete your profile.  

A second way: In the footer, under the heading “KLOUT FOR DEVELOPERS,” click on “Privacy,”  and you’ll go to step “4.” above.  (“Developers?”  Why only “Developers?” unless you’re deliberately trying to make the link hard to find?)

The third way: In the “Help” section, under “Your account” there are five help articles and then a link to “more.”  None of the five articles are about how to delete your account.  If you think to click on the “more” link you’ll come to a page of ten help articles, again none about deleting your account.  Finally if you click on page “2” you’ll see the final three articles, one of which is about deleting your account. 

Tricky, huh?

By contrast, to delete your Facebook account you go to Home / Account Settings / Security and there is a clear link at the top of the page that says: “Deactivate your account.” 

My point is if you brought 100 users to the Klout home page for a usability test and gave them the task to delete their account, I bet 90+ wouldn’t figure out how. 

It’s this kind of sneaky behavior plus the weasel words that undermine any message Joe Fernandez and Klout are trying to send about transparency and authenticity.

Ugh, Enough! 

It’s time to get back to work.  I’ve got better things to do. 

Going forward Klout will or will not implode as the forces of economics (and Google, I expect) work in the marketplace. 

I truly hope Klout goes away, but maybe it won't---maybe Kleiner Perkins will get their $100 million payday when Klout is sold.   Who knows?

But if you can’t figure out how to delete your Klout profile, then do yourself (and all of us) a favor and please, please at least ignore Klout. 

Go out and engage with interesting people on social networks and give out more than you receive.  Try to help build communities, and make them better and more valuable to everyone. 

You don’t need no stinking Klout score to tell you if you’re doing the right thing in social media.

rohnjaymiller

Rohn Jay Miller

Director of Digital Strategy, Hanley Wood Marketing

I'm Director of Digital Strategy for Hanley Wood Marketing in Minneapolis.  We've been in the content strategy and branded content business since 1984.  We deliver across content strategy and branded content. I speak in public, often by invitation. I can be reached at rohnjay@rohnjaymiller.com

See Full Profile >

Comments

tonyveroeven
Posted on November 29th 2011 at 4:14PM

"By contrast, to delete your Facebook account you go to Home / Account Settings / Security and there is a clear link at the top of the page that says: “Deactivate your account.”

Careful Rohn, about the Facebook Profilee activation Assumption--Check this out:

http://www.groovypost.com/howto/security/permanently-delete-your-facebook-profile-account/

Facebook has had the same complaints for years about deactivating vs. really deleting.

But point taken, neither want company want us to leave easily.

rohnjaymiller
Posted on November 29th 2011 at 5:19PM

Good point, Tony.  I think they all learned the "negative pick-up" strategy from GoDaddy and magazine subscriptions.

ilovegarick
Posted on November 29th 2011 at 6:07PM

Ok, if it's so blatantly evident that Klout is NOT the standard for measuring influence, then what's next? What do you say when no one else will admit that the emperor wears no clothes? Well, I've heard some good things about PeerIndex and Twitalyzer. Additionally, I'm interested to see how Kred works out too. 

And though I may not stake much value in Klout; I will be honest in saying that I was rather pleased to see that I qualified for a free Windows phone. Straight up, am I a sucker to be bought with the offer of a free phone? But then again, I'm not trying to "game" my score.. I simply talk via Twitter as a social channel. So would I be a sucker for NOT capitalizing on the chance to receive a free phone? 

 

 

MatthewColin
Posted on November 29th 2011 at 6:11PM

Sounds like you, along with many others, have misinterpreted the Cornell/Yahoo study.

From the abstract: "The result does not imply that 50% of tweets are broadcast by 20,000 users. In fact, the 20,000 “elite” users in question broadcast only a very small percentage of all tweets. However, many of these “elite” users have huge a large numbers of followers, thus their tweets constitute a much larger percentage of what other users receive."

Correct me if I'm wrong, but from what I understand, a better way to word it would be 50% of content (meaning links, hashtags, etc.) is generated by 20,000 users. Similar, but not the same.

rohnjaymiller
Posted on December 2nd 2011 at 5:07PM

I'm aware of the point you make.  My take of the survey is that in the sense that RT's are sending the same message as a Tweet, the statement that 50% of tweets are created by a core of 20,000 (very popular) Twitter accounts is true.  May be splitting hairs semantically, but I saw precisely the disclaimer you quote and thought awhile before writing it the way I did.  Thanks for the comment, though I disagree,

Andy Gonzalez
Posted on November 30th 2011 at 5:44PM

Great blog. You make some interesting points. I would completely argue against your request for Klout to make their algorithm public. Who does that? No one. Not Google, not Yahoo, not Bing, not Facebook not anyone who want to maintain the integrity of their platform. The problem with making the algorithm public is that you then open yourself up and lose any safeguard against people gaming the system.

rohnjaymiller
Posted on November 30th 2011 at 5:59PM

<server error>

rohnjaymiller
Posted on November 30th 2011 at 6:02PM

Andy: "Opening the algorithym" could be as simple as saying "Net followers + RTs + RT's of RTs = Influence."  My point is that without having to either disclose precisely how they calculate the numbers nor having their results audited, then Klout is free to say any damn thing they want about anyone they want.  Who knows if they're measuring "social influence" or what?  It's just their word.

Here's a big difference between Klout and Google:  if Goggle does a lousy job of ranking the value of Web pages, then people will stop coming to their site and use another search engine (which is exactly why Google did the massive overhaul of their algorithym last year to stop link farms.)  

If Klout does a lousy job of ranking human beings, what's the consequence?  How do we tell if they're doing a lousy job?  The answer is there is no way at all.

Google is a great search engine because when I type in my query I get really interesting, relevant results.

Klout is bogus because when it says you're a 41 and I'm a 51 there's no way of telling if they're right or wrong.  And no matter if they're right or wrong there's nothing you can do about it---and it might cost you a job if you're a 41 in a job applicant pool filled with 51s.

As I point out there are plenty of precendents for companies that do rankings to have those rankings audited and certified by third parties.  The fact that it's hard to do that in social media where so many people are trying to game the system is no exuse to give Klout a free pass.

Klout is a bad idea whose time has not yet come.

Andy Gonzalez
Posted on November 30th 2011 at 7:12PM

I guess we see things completely differently. I have a few points:

 

-I believe Klout is a great idea who's time is just right.

-I believe that there is an absolute way to justify someone with a 41 -vs- a 51.

-If you are being hired by a company who bases their hiring on a Klout score, then they aren't a company worth working for. 

-I think we all already know the "Net followers + RTs + RT's of RTs = Influence" formula as a Klout score in its basic form. Therefore, you have answered your own question. 

 

Opposing view points aside, I think your article was a good read. Hope you have a great night. 

 

Elias_Shams
Posted on December 1st 2011 at 5:22PM

Rohn, Excellent points! Your article is right along what Craig Newmark toldme last year when I was building my gig, http://awesomize.me  which I would love tosee your page on it and get your feedback.

I met Craig last year during the FedTalks 2010 event in downtown DC around  the time I was thinking of  building my next gig.  I managed to take 10 minutes of his time during the break to yap about  startup and the business model for the company. One minute after my yapping, like a good doctor, he gave me the good news and the bad news.

The bad news was that my idea of creating a directory of influential people was not quite original. He told me that there were other sites like Klout that have been working on a similar concept, but that they were all doing it using analytical methods. Then he gave me the good news.

The good news was that none of these sites have emerged as a winner in this space, and in fact he had doubts if any of them could survive. He especially liked the fact awesomize.me was trying to get the user engagement to rate other users as opposed to doing it analytically, as he felt that this model would work if we implement the right rating algorithm.

Then he gave me a dose of reality (more bad news?). He warned me that the major challenge with awesomize.me would be getting users to register on another social media site.

His advice triggered me to use the experience I had gained with my first startup, telezoo.com during the Web 1.0 era, and move forward in implementing the awesomize.me model for the enterprise. This way companies would certainly get their clients to register and rate their companies,  products and services on the platform. The decision allowed companies to create their page on awesomize.me attracted a decent number of companies to sign up  - about 1,500 of them including heavyweights such as AT&T, HP, Dell, CentruyLink, Honda, and The Huffington Post.

Well, Craig’s prediction about identifying the influential people  the awesomize.me way leveraging the real votes from the the community vs others like Klout that does it analytically was right.  The tweet below from Cute Culture Chick blog is one out of thousands that I’ve been spotting on the social media sites.

http://twitter.com/#!/cuteculturechic/statuses/138704164841717760

RobertFrench
Posted on December 1st 2011 at 5:32PM

You forgot the most important questions Klout cannot answer:  "Are you influential?" or "Do you have influence?"

Mark Cunningham
Posted on December 1st 2011 at 6:10PM

Thanks for the follow up.  I concur completely.  I'm one of the +90% you suspect who couldn't figure out how to delete/opt out.  Your post was a HUGE help.  I just opted out of Klout.  As required in the final step I was required to select a reason however my reason, "Can't stand their +K spammification by proxy and its gamey sour-Klout score zombies" was not listed...so I seleced "other" as my reason for leaving and wrote this candid note.

"The reason why I'm leaving is that I dislike the +K "spamming" that I perceived as Klout's self promotion by proxy.  Also, I no longer believe Klout has any objectivity.  It's my observation on Twitter that people are "gaming" Klout for scoring purposes and frankly, this utterly sours-Klout such that I no longer trust your score algorithms.  Lastly, I'm disheartened by the recent blog posts that have been very critical of Klout and Mr. Fernandez' handling of the overall scrutiny.   

Sincerely,  
@cunningham_mark (aka Mark Cunningham)

toniaries
Posted on December 1st 2011 at 8:24PM

Great job on these posts - this is a great summary of the unanswered issues, and your interview was informative, too.  Thanks!

LisaThorell
Posted on December 2nd 2011 at 12:38AM
Thank you for keeping these continuing technical issues with Klout alive. Even as much your post represents one of the most comprehensive summaries of Klout's deficiences to date, the company has recently received its largest injection of capital confidence yet http://read.bi/s4g2W9 And yet you raise one of the issues in Pt 3 that has deeply concerned me in a earlier post http://bit.ly/eKmCVb - namely the transparency issue. As you write: 3. How can anyone trust Klout scores when you keep your algorithms secret and you won’t allow an independent third party to audit their accuracy? How can you claim transparency when you’re not transparent? The reason we should worry here, folks, is that the latest capital injection is based on assuming Klout has the capability to be akin to a FICO score. However, even as much as you can look into the components underlying your FICO score, or see the components of a Black-Sholes model to understand your company's evaluation, or read the basis of your personal FICO score, you cannot view the underpinnings of Klout's algorithm judging your "online social influence":. It is not transparent in any manner. True science demands transparency. If Klout is to take these vast sums of capital injection as a future standard, we as consumers (with our "assigned" Klout scores) need to demand and be relentless in our demand for transparency in these algorithms. As you point out - for one company to be responsible for an online influence number, with no cross-checking, no auditing - not the simplest of scientific tests nor words from the influence networking experts who continue to remain silent on Klout- has the potential with this unchecked number to provide a baseless social caste bubble across many industries. The fact that people will lose job opportunities due to their Klout score, is a forward tragedy. The sad fact is that this is a marketing juggernaut, now powered by substantial VC funding, that continues to be unchecked by any rigorous scientific testing.
thelondoneer
Posted on December 6th 2011 at 8:46AM

The fact that Klout's scoring system is completely opaque troubles me. So does the fact that, in common with a lot of people, my score has been fluctuating in the 20+ point range for the last week gives me zero confidence that they actually know what they're doing.

I'm de-authorising Klout just in case - if someone wants to check my level of influence they can look at re-tweets, Google +1s, comments on my blog etc. Klout clearly isn't suitable as a form of social media shorthand...