Courtesy of Involver
Dear SEO consultants,
If you are taking a healthy dose of digital marketing every fortnight or so then you have probably heard of the new boys in town, yes the infamous Klout!
Klout has received a lot of good PR, if I can recall correctly they have been featured on New York Times, Mashable, Washington Post, ZDNet and I think they have been talked about on CNN as well (although I am not too sure).
Here is what Klout says about their scoring mechanism:
“The Klout Score measures influence based on your ability to drive action. Every time you create content or engage you influence others. The Klout Score uses data from social networks in order to measures your True Reach, Amplification and Network Score – you can find out more about all those three elements here.
Some brands are using Klout to measure user influence and based on certain thresholds of KScore they offer Perks, offers or in some cases coupon code on Facebook. A few months ago Klout teamed up with Involver to create Klout based applications for social media marketing on Facebook.
In their announcement Involver states:
“Brands facilitate this engagement by setting an appropriate threshold for how influential a user needs to be in order to receive their most valuable content. Users are rewarded for getting their Klout score but given an even greater reward if their score is high.”
Involver describes their Klout Coupons application as follows:
“The Klout Coupons App from Involver is an innovative way to engage users while targeting your most valuable promotions to your most influential fans. Klout is the leading way to measure online influence based on Facebook, Twitter and LinkedIn profiles.”
Big brands like Audi have used Klout Coupons, the question is, do they really know how accurate Klout is? Can it be relied upon to measure someone’s social influence?
Is Klout’s scoring model based on complex scientific analysis or is it just a simple mechanism which can be influenced through simple manipulation? Well, some of your might argue that it is scientific, accurate and so on.
Personally, I don’t buy all the hype and the marketing talk but I do like their PR agency, whoever it is, they are doing a good job – so far.
Anyway, it is really late at night or morning should I say, 1:42am to be exact. Let me introduce you to @_BorgCollective and shed some light on Klout’s scoring model.
@_BorgCollective follows 46 Twitter profiles, has 104 followers and is listed 6 times (as of 22/09/2011 at 01:43am). Wonder what its Klout score is?
A whopping 57, see screenshot below:
http://klout.com/#/_BorgCollective
So http://klout.com/#/_BorgCollective scores 57 out of 100 which means (S)he is more influential than me, by two points so far. At this point you should be asking yourself – how did (S)he manage to get that score, right?
Now based on Klout’s advice, you and I should do the following to increase our Klout score:
- Create content worth sharing
- Start discussions
- Register and connect your networks (Latest being Google+)
As you can see in the screenshots below, @_BorgCollective’s klout has been consistently increasing in the past couple of weeks.
Score Analysis
So everything looks fine, @_BorgCollective’s profile is getting some juice, (s)he must be really engaging with some influential people to have that steady increase, right?
Klout Fails
What does @_BorgCollective talk about and who does (s)he or it engage with? The account is only connected via Twitter and no other networks so Klout can only assess its influence level based on Twitter’s data. Problem is, @_BorgCollective doesn’t really engage with anyone, it doesn’t really “influence” anyone but it has a higher score than me. In fact, all (s)he or it does is tweet the same thing over again and again – basically spamming other Twitter profiles!
As you can see below, the entire timeline is full of @someusername “YOU WILL BE ASSIMILATED”:
_BorgCollective’s public timeline on 22/09/2011
How to do you game Klout?
Look at what @_BorgCollective has done, that profile is simply mentioning random people with the exact same message three to four times every hour. It is not connected to any other networks, no profile information and no external links.
Should you invest in Klout based campaigns?
No. How can a brand like Audi depend on such a flawed analysis and scoring model? How could other marketers who are currently offering perks on Klout.com rely on such a flawed and what seems to be easily game able system?
I have been monitoring Klout and Peerindex for a while now and I have always thought that Peerindex is more realistic in terms of its analysis. Now even though I believe Peerindex is slightly more accurate in some ways I wouldn’t buy things at face value purely because all these technologies are very new. With that said, personally I have developed a tool that fetches me Klout, Peerindex, Followerwonk, looks into some other stuff and gives me a mean score. I think it is slightly more trustworthy because I am getting data from several scoring systems. That is my automatic way of doing it, I highly recommend carrying out influence research manually if you are planning on a big budget campaign.
What is with sudden Klout score spikes?
Here is another anomaly, the CEO and co-founder of Klout, Joe Fernandez scores 72 out of 100. But here is an interesting fact, on August 24th he scored 68.47 and suddenly on 28th August keep jumps to 73.06 out of 100, see blow.
I will be honest, tonight I have not had a chance to investigate why Joe Fernandez’s score shot up so quickly. It could have been as a result of a PR campaign or something similar but taking the http://klout.com/#/_BorgCollective/ scenario into account I think Klout is not and can not be “the standard measure of influence”.
Now that you are here, it can only mean two things, you either want to share this post on Twitter or you are interested in reading Google: The Small Business Serial Killer.
UPDATE: Made some more shocking discoveries about Klout, check my new post.
Do you think that Follwerwonk and Peerindex do not have the same issues or are more trustworthy algorithms?
Hi Mark,
All three of them have issues, bear in mind if any of these were super hot start ups then Google would have gone out of its way to acquire them for their authorship analysis.
With that said, I think Peerindex is slightly better than Klout purely because their scores don’t bounce up and down sporadically every day.
Yousaf – thanks for the feedback and appreciate your thoughts. In terms of Klout’s changes, I might suggest that Klout (as well as PI and FW) might be tweaking their algorithms to handle new data into their influence algorithms. I can tell you were are regularly tweaking our formulas to better represent your social capital.
And yes, we are at the cusp of a new solution that will give us some new insights. And looking at exhaust data as well as intent will help us better understand our personal influence on the world.
Thanks very much for sharing – one thing that I think is worth noting as well is the impact of one or two big tweets that get retweeted a lot (a joke or whatever) I know that this can have a massive impact and may explain the spike on Joe’s account. Alternatively, I’ve also seen spikes like this when people add a new profile/account from another social media platform so maybe he’s got a lot of friends on Facebook or something.
Great job in highlighting some of this stuff and for sharing specifics though I must say I came here from a tweet that said “Amazing how many people take Klout seriously when sh*t like this is happening” and I was a bit surprised, does anyone actually take this seriously? I thought it was pretty well known that it’s quite easy to game and not a particularly meaningful number?
Hey Sam
You are absolutely right about one or two tweets having a sudden impact and yes connecting to a new network does cause spikes as well. One of my colleagues linked his Facebook account and within a day his Klout jumped from 15 to 36, two weeks later he ended up back to 17.
I think these knee-jerk fluctuation are very stupid purely because influence can only be determined over a long period of time based on a large data sample. Yes there can be some exceptions i.e. if something went viral – like really viral!
Me and you don’t take it seriously Sam, but Audi, Starbucks, Coca Cola, Cover Girl, Spotify, RedBull, Subway, Chevrolet etc.. do take it else they wouldn’t have launched Klout based campaigns?
No idea, who persuades them to do so but I think most brands are so confused about social media that they are ready to try anything just to see what works.
I normally do all my outreach research manually but as I have mentioned in my post I do use my own tool sometimes.
I think that with companies like coca-cola, audi, ect they are just trying to play catch up in so many areas of social marketing/networking that there willing to just try everything and see what works and go from there really. I’d imagine that these large companies could and would very easily hire consultants that have more expertise in this area so they have a more focused approach because it seems like they believe whatever they are told. A lot of data can be skewed one way or another depending on how you interpret it.
Cool, thanks for the feedback mate. As I say it’s good to see the practical “this is what impacts scores” kind of posts – as for the Audi, Starbucks, etc. doing Klout based campaigns that is news to me and quite shocking news if I’m honest.
As with any other metric I suppose Klout score is as good a starting point as any (saves some legwork if you are looking for influencers) but I guess much with everything else these days you still need the manual side of things!
Cheers again for sharing,
Sam
Hey Yousaf,
Thanks for taking the time to check out what we are doing here at Klout.
We believe influence is the ability to drive action. Is this BorgCollective account silly? Absolutely. Take a look at twitter search though (http://twitter.com/#!/search/%40_borgcollective) and you will see a long list of influencers engaging with that account over the last few days. It’s a meme that is influencing people right now and will likely die down quickly.
You also mentioned the big score jump on my account in late August. I remember being surprised at the time by it also but in looking in the data found I had a single tweet retweeted more than 100 times including by massive influencers like @kevinrose (http://twitter.com/#!/JoeFernandez/status/106793901888770049).
The fact that a joke tweet from me or a silly sci-fi reference generates actions and engagement doesn’t suggest our system is any more flawed then Jersey Shore being one of the highest rated TV shows. We can’t be blamed for people’s taste.
All that said, I do appreciate what you have pointed out here and we are constantly improving our algorithms. Our science team is going to look at this account and in fact we are on the verge of releasing some pretty major changes around how we score.
thanks!
See! Told you [fill in the company] change their algorithms all the time!
That reminds me of Google’s “500 improvements a year”.
Hi Joe,
Thanks for your comment. It goes on to show that you care about your product and more importantly about its users.
Your definition of influence does resonates with me, it really does. We both agree that BorgCollective is a rather silly account. I like to think that I am a scientist, so what I always yearn for is accurate in-depth analysis instead of slogans. If I paid you £10k and found out that a spambot can reach a score of 57 within 30 days, I would want my money back. That is just me, Audi can afford not to do that.
I am not trying to single out BorgCollection’s case, my concern actually stems from elsewhere. Stating that Klout is “the standard for influence” is a rather lofty aspiration, something that is cited often but never really sighted – at least from a scientific perspective.
I think there should be a clear distinction between “noise”, “activity” and “influence”. If someone is active and noisy, it does not necessarily mean that they are influential. From what I have seen, I think Klout is simply measuring levels of activity; without filtering out noise and is therefore not in a position to measure influence.
Correct me if I am wrong but I think Klout and most of its competitors do nothing but primitive analysis. From what I can tell all you are doing is looking into Twitter activity, retweets and basic profile signals etc (at least in BorgCollective’s case).
The second thing I would like to bring to your attention. Let’s say I am a very “influential” guy currently scoring 70 out of 100 for “physics” and “chemistry”. For one reason or the other, I don’t have access to internet for 2 months. When I am finally online, I discover that my score has dropped to 15. Does that mean that I know less about “physics” and “chemistry” or that I am no longer an authoritative voice for those subjects because my level of activity dipped for a couple of months? Do you see what I am trying to get to?
On Klout, scores go up and down almost every day. Levels of activity can fluctuate like that but surely influence can not be that volatile, no?
With that said please bear in mind that I do understand some of the challenges you and every one of us face when it comes to social web. Everyone needs to start from somewhere and we never get it right the first time.
And yes I do understand, you are in it, for me it is easier said than done.
I would have thought having 1,861 identical tweets would have raised a flag somewhere.
Nice article, but since when has Twitter been about “scoring”? I thought it was about social interaction?
It is about social interaction but similar to real-world we the marketers want to find out who the movers and shakers are based on, let’s say, subjects, market sectors so on and so forth. Klout & Peerindex are trying to detect and score influential social profiles to help us with our outreach. Twitter itself is not about scoring, but we as marketers need data to ascertain impact and propagation.
Really great to see Joes comment on this! At least this shows he is not like other startups that make their product and forget about who is earning them the money (user based products).
Like Sam said, I was also surprised that people are really taking Klout score seriously? I mean I do check it when I have nothing to do anything else but that’s just to bring a good smile on my face and that’s it!
After reading Joe’s comment, well I can wait for a better and less game-able Klout but for now investing on basis of klout like Audi did will not be a very good idea…
Good work Yousaf mate!
Moosa, thanks for your comment! Good too see me, you & Sam agree on this!
LOVE this article. Completely agree with you. No sh!t, black boxes can be gamed…
Hey Robert,
I am glad you like it, I just posted another one you might be interested in checking it out http://www.rocketmill.co.uk/gaming-klout-second-casestudy
[...] 24, 2011 by Yousaf Sekander under SEO TweetTweetLast week I posted an article highlighting how flawed and easily manipulatable Klout’s scoring algorithm is. In my post and the subsequent commentary I argued that:Klout model of scoring seems to be based [...]
Hey Yousaf, great article and your very right about Klout and hats off to your investigative skills, i must say i use klout to measure how i am progressing with social media, but there are bound to be ways to game Klout, it must be said that Klout is still a beta and as a result of this its measurement of Klout score will be much more premitive than it will be in the future. Also by using social media platforms alone it must be much more difficult for it to get an exact measurement because of the limited information that is displayed, especially if you compare a Klout score to Googles page rank,
[...] just read this post on gaming klout, and this follow up that includes more examples of gaming klout. And my heart [...]
[...] or report the breaches to Massachusetts regulators." Why is the identity theft Apple's fault?Here is how you can game Klout >> ElevateLocalSurprisingly easily, it turns out.Exclusive: four pages from The Zen Of Steve Jobs >> [...]
[...] Reflects. Only Steve. Presidents Remarks. One More Thing! Google Class Act. Jonathan Mac Image. Gaming Klout 100 #1 Gaming Klout 101 #2 Microsoft courting Yahoo again? Cloning Stem Cells. Pirate Bay lives on. Delta [...]
[...] just read this post on gaming klout, and this follow up that includes more examples of gaming klout. And my heart [...]
[...] recent tweet from @LouWoodley alerted me to a post which described “Here is how you can game Klout“. The post described how an automated bot seems to have been successful in gaining a high [...]
[...] on levels of network activity which is a rather primitive appr…Posted in Google Read More Klout: Here is how you can game KloutIf you are taking a healthy dose of social media every fortnight or so then you have probably heard [...]
[...] on levels of network activity which is a rather primitive appr…Posted in Google Read More Klout: Here is how you can game KloutIf you are taking a healthy dose of social media every fortnight or so then you have probably heard [...]
[...] ????????? ??? game Klout ??? ElevateLocal [...]
[...] Internet community has mixed opinions on Klout. There are individuals that have figured out how to game the system. Also, there are individuals that are greatly upset when Klout updates the algorithm and [...]
[...] about on Twitter or Google+ all day or even better – a bot lol ….and pretty easy to game Klout apparently [...]
[...] on levels of network activity which is a rather primitive appr…Posted in Google Read More Klout: Here is how you can game KloutDear SEO consultants, If you are taking a healthy dose of digital marketing every fortnight or so [...]
[...] Reflects. Only Steve. Presidents Remarks. One More Thing! Google Class Act. Jonathan Mac Image. Gaming Klout 100 #1 Gaming Klout 101 #2 Microsoft courting Yahoo again? Cloning Stem Cells. Pirate Bay lives on. Delta [...]
[...] Dico trucchi, perché l’algoritmo segreto si può un po’ prendere in giro, come ci dice questo articolo. Che poi, il funzionamento di questo fantomatico algoritmo non è ben chiaro a nessuno fino in [...]