Online influence tools carry no klout with me

Michael Litman says online reputation tools are fatally flawed and reckons it’s time to get back to good, old-fashioned interpersonal relationships.

The first time I heard about Farmville, which generated $235 million in revenue in its first quarter, I got it but I didn’t really get it. A similar thing is happening now with the onslaught of new online influence tools.

I’ve been writing about online influence services since 2008. During that time, companies like PeerIndex, EmpireAvenue, Klout and some newer varieties like Kred.ly and Proliphic have sprouted up, seeking to try and solve the problem. I don’t think any of them have done that just yet.

They all set out to, in some way, shape or form, to define who is “influential”. But do we really want everything we stand for as people to be reduced to one number that only credits us with the extensiveness or lack thereof of our online activity?

Let’s get this out of the way now: the reputation graph seems to be just beginning to gather steam. Klout recently announced they have raised a C round of financing, purported to be around $30 million, and they topped 10 billion API calls in December 2011 alone, up from 100 million in January 2011. (Though half of that was probably from Robert Scoble.)

Who cares?

There is only one person on the internet with a perfect Klout score. Yes, I’m afraid it’s Justin Bieber. Bieber is more influential than the President of the most powerful country in the world, Oprah and even our esteemed editor. Proof, you might think, of the depressing meaninglessness of it all.

We all hold some form of influence over friends and industry peers, whether it is knowledge about great places to go for food, the ability to recommend a book or simply the fact that we share interesting links. Automated influence tools will never be able to pick up on these subtleties, though, because they only look at aggregated, descriptive data from our online activity to predict outcomes.

Phil Guest, chief commercial officer of PeerIndex, explains how it can go wrong. “We call it the ‘Clay Shirky problem’, he says, “Where Clay for his level of influence in defining market sectors, is not very active on the social web and as such has a relatively low ranking.”

There are three types of measures that social tools can use: descriptive, diagnostic, and predictive. Descriptive data is what has happened, diagnostic data tells us why and predictive data helps to figure out the future, based on the present. Influence tools are descriptive measures, they tell us what’s happened – but that’s it.

That’s the short version of why we shouldn’t get so hung up on the number these tools give out, absent any other information. Investigating and cultivating the human side of influence is required for the rest: doing it by hand, to build long-lasting relationships.

Good sales and marketing is about building trust over time, and thus selling more ‘stuff’ to the right people at the right time. There’s no point contacting an influential food blogger with an unrelated technology product. It’s a waste of your time – and, more importantly theirs. They’ll probably never read another of your emails, even if you have go on to present something of relevance later.

With deadlines getting shorter and more pressure being put upon on the advertising, marketing and public relations industries, online reputation tools are purported to be the holy grail of “quick fix influencer discovery”. But, in truth, they barely scratch the surface for you.

While using them may bring to the surface the people that the tools think are influential, these influencers still won’t know who you are, why you are pitching to them and why they should care. No automated tool will ever help you to build real relationships.

That’s why I can’t for the life of me understand why Klout just got $30 million dollars in funding. What, exactly, are the VCs investing in? John Frankel of ff Venture Capital says: “The reason we invested was that we believed by mid-decade people would care about their online reputations and would need tools to manage it.” Hmm.

Who, what, how

The methodology for services like Klout is famously kept a secret. That pisses off people when they see their score dramatically plummeting overnight because of a tweaking of the magic formula.

It infuriates people because they see crazy stories about companies who screen potential employees based on influence scores and they see others adding Klout scores to their resumes. This is so unbelievably wrong-headed. Klout will not improve your existing relationships with people, nor will it make you better at your job or a better hire.

And what about the “topics” Klout tells us we are influential in? They are often more hilarious than they are accurate. Go have a look for yourself. I’m influential in Sony Ericsson, Red Bull and Gowalla, apparently – three things I don’t generally talk about or use.

What really is the fascination with classifying people as a number online when the methodology behind that number is unknown and so can never be universal and useful? We saw how a corner of the internet was in a state of shock when Klout tweaked their formula recently.

One person commented: “I am beyond irritated over my 10 point drop, plus demotion in title! I work in Social Media Marketing, this point drop will hurt me in gaining clients! What the hell!?!?!?”

Another said: “This trashed a 6 month effort to get our organisation to use Klout as a measure of social media marketing effectiveness!”

This guy is apparently being serious: “My score went from 73 down to 53. 20 point drop. I’ve been working for months to increase my Klout score. Please fix this.”

Only Klout know why there are such wild fluctuations in scores. As individual users, we’ll never know why and what to do to remedy those changes. In other words, today’s influence tools are not diagnostic. They don’t tell us why things have happened.

So Klout eroded transparency and trust by tweaking people’s scores without explanation. It doesn’t take a psychologist to work out that no one likes their perceived worth to diminish overnight for no reason.

Phil Guest says that’s where the differences between Klout and PeerIndex become more pronounced. “Klout gives you free stuff. Our approach is that the cool stuff goes to those that matter to the brand, and is seen to be provided by the brand, not PeerIndex – we’re just the engine. This is important: the brand [is better] able to build a long-lasting relationship with a consumer.”

I also spoke to CharlotteMcEleny, a senior journalist at NewMediaAge says, “There is a real need to find a measure for influence. Social media has been blighted by the issue of measurement.

“The industry blindly follows some of these measures without bothering to understand the methodologies behind them. All forms of marketing measurement methodologies hold flaws, the key is to understand those flaws. It is also up to these companies to retain credibility by maintaining transparency over their methodologies, or they risk becoming redundant due to user distrust.”

Could Google fix this mess?

With its recent foray in to social layering across the entire internet with Google+, and its history with PageRank, Google has more experience measuring influence than anyone else. The company is now well placed to take those learnings and build another set of data to add to their arsenal: social influence scores.

Vice President of Product for Google+, Bradley Horowitz reminds people to not see Google+ as a standalone product, but instead as an assault on Google itself to become a socially focussed company from top to bottom.

It is, as it ever has been, ultimately all about relationships and people. These things are manual, long-term processes that require time and expertise to yield maximum benefit.

So I’d like to see us getting back to the human side of influence and taking control back from tools and algorithms. No automated tool will be able to establish long-lasting personal work relationships like you can.

Michael Litman writes for The Kernel on social media. His Klout score is 49.