Klout’s bunk – Everybody can be an influencer
Last week we showed why Klout not only does not measure Influence directly, and we also argued that it also does not measure social capital as Brian Solis claimed for Altimeter in a recent report.
Now, whether Klout, PeerIndex and Kred measures social capital or status (as we believe it does), most marketeers are interested in the bottom line. Are these tools a predictor of a person’s capability to influence action?
This is where it gets interesting. As Philip Sheldrake explains – influencing action in is the realm of complexity science:
Think about it. Take almost any of your recent thoughts or actions and try and decipher how in fact that thought or action came to be; what did you take into account, consciously and unconsciously, over what timescale? You soon begin to appreciate that your thoughts and actions are outputs of a complex system. You are reconciling multiple inputs, multiple influences.
In other words, it’s not that easy to say whether a person will indeed influence action.
You might – as many marketeers do – vehemently disagree. You know in your gut that if you get Stephen Fry to tweet about your promotion, thousands will see it, and if its a compelling message many will retweet it to their followers. There will be some brand uplift, and with a bit of luck, a measurable rise in sales. Simple.
In a recent excellent Forbes article two people conveniently articulate either side of this debate really well.
In the red corner, we have Malcolm – social media can’t start revolutions – Gladwell himself. In his book The Tipping Point, he argued quite convincingly that certain people (taste makers) are more influential, and can start trends, that spread to the general population. His thinking is neatly encapsulated in this quote:
“in a given process or system, some people matter more than others.”
Gladwell used the famous 6 degrees of separation experiment to prove his point. 160 letters had been given to individuals in Nebraska, with the task of giving them to someone closer to a stock broker in Boston. Besides the fact that it generally took six turns, what excited Gladwell was that half the letters reached their target through only three people. There were his superconnectors.
In the blue corner we have Duncan Watts. An Australian scientist who has spent a large part of his career trying to make sense of Influence. Watts did not think this ‘gatekeeper’ model was true and conducted computer simulated experiments to prove it.
In 2001, Watts used a Web site to recruit about 61,000 people, then asked them to ferry messages to 18 targets worldwide. Sure enough, he found that Milgram was right: The average length of the chain was roughly six links. But when he examined these pathways, he found that “hubs”–highly connected people–weren’t crucial. Sure, they existed. But only 5% of the email messages passed through one of these superconnectors. The rest of the messages moved through society in much more democratic paths, zipping from one weakly connected individual to another, until they arrived at the target.
Watts had proved that Milgram’s (the guy behind the 6 degrees experiment) sample size, when it came to Gladwell’s superconnectors assertion, where too small. In other words, when it comes to spreading information we all matter.
Why was Gladwell (and many others before him) wrong? Gladwell argued that trends spread like viruses. But says Watts, different viruses spread quite differently, and that those arguing for an Influentials theory do not put forward these nuances:
“…there are a lot of ways an Influential could convert the masses. Merely talking to a friend once could infect her with an idea. Or it might take several conversations. Or maybe Influentials are so persuasive they’re like trend vampires, and each victim they bite becomes hyperpersuasive too. Depending on how you define the specific mechanics of influence, you’d get totally different types of epidemics–or maybe none at all. But gurus of the Influentials theory never directly clarify these mechanics.“
Watts conducted another experiment. First he created influentials that could ‘infect’ four times as many people as the average person. The result was fascinating, because even when he increased the Influential’s infection capability to 40 times that of ordinary people, he found that they were not that special. Trends started by influentials, spread further once they got going, but average people were still almost as likely to start a trend.
Why could the influential not kick off a trend every time as some would expect? Watts believes that when starting a trend it’s not that important who starts it. What’s far more important is societies readiness for that trend. Large cascades (as social scientists call the spread of information) occurred when easily influenced individuals influence easily influenced individuals, and not when “influentials are influencing followers”. In other word’s if the conditions are right, almost anybody could start a trend. Trends are more like forest fires than likes viruses.
Watts has a third experiment where he pushes this proposition even further: Trends are not cultural certainties, but rather trends are actually quite random.
That may sound bonkers, but this is how he went about proving that:
Watts built a world populated with real live music fans picking real music, then hit rewind, over and over again. Working with two colleagues, Watts designed an online music-downloading service. They filled it with 48 songs by new, unknown, and unsigned bands. Then they recruited roughly 14,000 people to log in. Some were asked to rank the songs based on their own personal preference, without regard to what other people thought. They were picking songs purely on each song’s merit.
But the other participants were put into eight groups that had “social influence”: Each could see how other members of the group were ranking the songs.
Where ‘social proof’ – ie. you could see how others had voted – snuck in, the results differed significantly from the world where social signals were absent. In the later the results were consistent, with the rank of songs showing little difference from one case to the next. In the former swings in popularity were dramatic, with songs that broke out as most popular in one instance doing badly in the next.
Influence makes for unpredictability
In other words being aware of the choices of influential users, not only amplified trends, but also made them less consistent and meritocratic!
Now this should make alarm bells go – if you are one of those people, who like I, hope that the Internet can play a positive roll in making our society a better place. But it should make your typical marketeer worried as well, as social influence and status seems to distort action in unpredictable ways.
Now – we have seen some evidence that what Watts is saying is true. Take the social media buzz around the death of Osma Bin Laden. Socialflow did an excellent analysis of how the story broke online.
So I’m told by a reputable person they have killed Osama Bin Laden. Hot damn.
— Keith Urbahn (@keithurbahn) May 2, 2011
Accidents of circumstances
Now if you don’t remember. News of Bin Laden’s death broke on Twitter first. And it was sparked off by Keith Urbahn, an ex Donald Rumsveld staffer, and a person with barely 1000 Twitter followers.
Socialflow pointed out that others with much larger followings had speculated that Osama Bin Laden might have been killed earlier, but the uncertainty in their language might have contributed to the information not spreading.
Before May 1st, not even the smartest of machine learning algorithms could have predicted Keith Urbahn’s online relevancy score, or his potential to spark an incredibly viral information flow. While politicos “in the know” certainly knew him or of him, his previous interactions and size and nature of his social graph did little to reflect his potential to generate thousands of people’s willingness to trust within a matter of minutes.
While connections, authority, trust and persuasiveness play a key role in influencing others, they are only part of a complex set of dynamics that affect people’s perception of a person, a piece of information or a product. Timing, initiating a network effect at the right time, and frankly, a dash of pure luck matter equally.
The result of this confluence of circumstances were quite something:
The rate at which Keith’s message spread was staggering. Within a minute, more than 80 people had already reposted the message, including the NYTimes reporter Brian Stelter.
This analysis from Socialflow chimes rather nicely with Watts’s 2006 paper which concludes that a network view of influence would show us that people “who later seem influential, actually were simply accidents of circumstances”.
So Klout is bunk?
Which brings us back to Klout scores and that Altimeter report.
Last week we said while Altimeters’s first two pillars of influence – reach & relevance – actually measures status, the third – Resonance – is a case by case proxy for actual influence on social media, but after the fact.
Resonance if you remember looks at how your messages actually travel in social media.
In other words, counting the Retweets and Comments a person generates is measuring a type of influence, but always after it happens. And the prevalence of this fluctuates wildly.
You regularly see people complaining that their Klout score fluctuates. But my contention, based on Watts’s research, is that if Klout made use of the Resonance pillar, it does not fluctuate nearly enough. One can only presume that systems like PeerIndex and Klout give less importance to the last pillar (and more to the other two) AND has probably also used some dampening algorithm: They have built a bit of a circumstances shock absorber to smooth over the rough edges of the fluctuations in influence.
So is that the last word? Klout can measure influence only after the fact, but they give preference to the other Pillars. And did Watts win the debate? Are Influentials not that special?
No, in a final twist in my next post I will show why many marketeers gut feeling on this subject is actually correct. It’s not that Watts was wrong, but he was not looking at the whole picture.
Wessel is the founder of RAAK Social Media.