Experts differ on Klout's clout

Popular social media "influence" firm Klout, took some knocks this week from folks angry about its changes.

The San Francisco-based firm produces a "Klout Score", from 1 to 100, indicating the (you guessed it) clout of folks across the social media talk-o-verse of Twitter, Facebook and sundry settings for our current era of computer-assisted solipsism. "The Klout Score measures influence based on your ability to drive action," says the firm.

But due to a tweak this week in the formula the firm uses to calculate its scores, some folks saw their numbers drop 10 to 20 points in a widely-lamented leap.

"The problem is that Klout's new scoring method is neither more accurate nor more transparent," said Brand&Capture's John McTigue, in a commentary critical of the change. "They have, in fact, made a serious mistake in attempting to redefine influence and may soon be paying a price..."

Well who cares? Some complained that the scores may affect decisions by employers on whether to hire someone, making the changes a serious matter.

So, we asked experts who have published on the topic, to comment on whether Klout really had any clout. We started with Morad Benyoucef of Canada's University of Ottawa:

Q: Is there statistical legitimacy to Klout measures?

A: I think Klout takes a rather simplistic view of influence. They say they measure how much a user's content is "shared" and "commented" on. They consider the topics being discussed and consider the user's network of friends and followers. Probably a fine algorithm but it does not measure influence. It measures how much a message with certain keywords travels, where it goes, who initiates it, who resends it and who comments on it, etc.

There is influence when the actions of an individual can induce his friends/followers to behave like him, e.g., in adopting a new technology, in choosing a brand over another, or in voting for a political candidate. Hence if I share a post or comment on it, it does not necessarily mean I have been influenced by the person who posted it. My comment could be contrary to what the initiator of the post intended. Or I can decide to share it with someone because I believe he might be interested, even though I have no interest in it whatsoever.

I say it is a simplistic take on influence because of two other points:

(1) Users in a network are not influenced in the same way. They are not identical with regard to the probability that they adopt a certain behavior. For instance, a user's decision to purchase a product might depend on the proportion of other users who have purchased that product. That proportion is the threshold, and it is different from user to user.

(2) A user's amount of influence on another user may depend on the relationship between them. My friend will influence me more that someone I do not know.

Q: Are serious businesses really using the scores for any decisions?

A: What kind of decisions? Hiring someone based on their Klout score? Making design decisions based on what influential people are saying? I don't know.

Klout provides "perks" offered by "brands" to influential users writing about them. Now what company would provide perks (and pay a commission to Klout) to have someone create content about its brand and push it through the network? Any company that believes in "viral marketing" would want to do that. But if it does it through Klout, it would be paying for 'impressions' not influence.

Beyoucef notes Klout's own views:

How can brands use Klout?

Brands can run Klout Perks and use Klout to measure their own success on social media. Klout Perks enable brands to reach and engage with influencers around a Perk. Klout influencers create thousands of pieces of UGC and millions of impressions for a brand's new product, initiative or campaign. Perks are targeted based on Score, influential topics, location and more. Your brand's Klout Score indicates how well you are driving action for your company. Monitor your Score to understand which tactics are working and how to better engage your audience.

"They clearly say that they measure 'impressions'. So Klout (at least for now) is another way for companies to get the conversation started around their products, and to keep that conversation alive, with the added value that it is done by users," Beyoucef says, by email.

Computer and information systems expert Kevin Curran of Northern Ireland's University of Ulster was positive in his views:

Klout is a revolution. There is 'gasping need' out there for a tool which can somehow gauge a person's 'power' in cyberspace. Whether Klout becomes the 'gold standard for influence ranking' remains to be seen. The concept is brilliant. Someone was always going to discover it. Will Klout be the one to perfect it? We will know pretty soon as first come - first wins in cyberspace.

To truly compare and rank the short infrequent 'bursts' from us an individual is as difficult a task to perfect as it is to get a machine to truly pass the Turing test. The main problem here is deciphering between the noise, activity and actual influence. Just because an individual is active and noisy, does not make them influential. The post that their followers are re-posting could be utterly useless to a large segment of society whilst the truly influential post may not contribute to their Klout rank.

(Q: Is there statistical legitimacy to Klout measures?)

A: At the moment, it seems that Klout does a primitive analysis. They mostly Twitter and Facebook activity, retweets, likes, list memberships, unique mentioners and other basic profile signals. The fluctuating score seems to be the source of much sorrow among Klouters. Klout themselves state that influence is relative so it depends on your goal and peers. In fact, the average Klout score is not 50 but rather around 20. Klout claim(s) that the score becomes exponentially harder to increase as we move up the scale so a move from 80 to 85 is much harder than 15 to 20. Klout stated lately that a majority of users would see their scores stay the same or go up but some users would see a drop. They claim this was necessary as their goal is accuracy above all else. That of course is a noble goal. A legitimate question here is whether 'an imperfect measure is better than no measure at all?'. Akin to the Coca Cola secret formula and the Google Page Rank Algorithm - Klout unfortunately does not release the intimate details of their algorithm.

Klout is for individuals who know how to influence and attract people. In truth, this is a skill that not everyone possesses. The truly great bloggers always come across as people who never sleep. In reality, they probably do sleep very little but they simply know how to continue to feed their audiences and encourage growth. Klout attempts to measure this activity and introduce these 'Princes of Cyberspace' to the 'Captains of Industry'.

The need for a Klout does exist. Marketers simply need information so as to ascertain impact and propagation of their products.

The Klout Perks aspect is truly a potential game-changer in marketing. Using Klout to measure their own success on social media is powerful. Introducing key businesses to the influential individuals in their market segment is a mouth-watering prospect. To some degree, so what if Klout have a certain percentage of 'misses' (i.e. including 'dumb' individuals with 'no real influence' in their lists). Do you not think that over time - these individuals will not be rooted out as a result of the continuous intelligent monitoring of ROI? That is important. The longer Klout exists - the stronger its core algorithm should become. Of course, the race is on to become the defacto measurement tool for behaviour (who would now think they can knock eBay or Amazon from Number 1 position?)

Klout perks in some ways (apart from the actual kudos of having a high score) are the currency of Klout. These perks are aimed at the individuals who obtain a certain threshold in their score. Of course, just because one obtains a very high score does not lead to that person being eligible for all perks. No, perks are offered according to location (located based advertising) and actual topics that they are classified as being influential in. Interestingly, the influencers who take the free flight are not be required to tweet or blog about their perk. Of course, they hope you do but it is not a legal obligation. (I wonder if they will track this in future and avoid offering perks to those who don't say thanks!)

Brands like this as they connect with influencers in their area of expertise. Perks offered to date have been along the lines of tickets, test drives, meals, gadgets. We should not also negate that fact that Klout is also striving to become a reputation system. Reputation systems are extremely valuable. Can you imagine eBay without its reputation system?

Klout also encourages a class system. One of the first online actually. To become ranked higher, one should actually engage only with people who have a relatively good to a higher Klout score. Engaging with people on Klout with a poor score is to be avoided at all costs......

Spamming to cheat Klout has existed from day 2 but the honest way to climb the ladder is simply to create content worth sharing, start discussions and register and connect your networks.

We all care about reputation, and Klout is one of the most prevalent ways to specifically measure personal online reputation currently (others include peerindex). I have a large respect for Klout as they are a first mover in an industry sector that will inevitably create much value.

(Q: Are serious businesses really using the scores for any decisions?)

A: Are serious businesses using it? Yes, but they are all basically dipping their toes in the water. They are 'casting their bread upon the water'. The reputation system is far from perfected. To a large degree, the only way it can be perfected is by business running trials and then allowing Klout to analyse the outcomes. The network effect is critical for all social networks so the days ahead will show whether the critical mass occurs so that the perks can flow. Again, the perks will flow when the brands arrive. The brands arrive when the influencers arrive. The influencers arrive when the .........

There is an slightly addictive quality to a social network ranking site like Klout. The ability to see a score associated with your 'influence' and being able to compare with your 'friends' does spur one on to increase it. This seems to lead to more online activity - heartfelt or not - but every post counts! In fact, if Klout did not invent Klout - then I there say that Twitter or Facebook would have invented it.....

That seems like about all there is to say about Klout for now. Hopefully, a Tweet brought you here. That will be sure to improve the ol' Klout score.