Michael Wu holds the exalted title of Principal Scientist of Analytics at Lithium, the SaaS provider of Social CRM services.  Last week he wrote a thoughtful piece about social influence that made the point that social influence scores are inherently false measures because they can be gamed—much more so than search engine results, for example.

There is a far more important conclusion that hovers over all of the discussion about social influence. Let me illustrate by pointing you to Nate Silver, the newly crowned king of political polling, who writes the very influential 538 Blog at the New York Times.  

As you may know Silver predicted on the 538 blog the outcome of the presidential election for all 50 states correctly.  Even Florida, which went for President Obama by fewer than 80,000 votes out of 8.3 million cast in that state. 

Trust me, Nate Silver knows how to predict elections.

How does he do it?  Silver is the accomplished master of aggregating polling information from a wide range of credible polling firms and then adjusting them for bias, and integrating them into a sophisticated mathematical model he has built which accounts for the other major influences on the election results such as jobless rate, growth of US GDP, and others.

If Nate Silver speaks about an election, people listen. 

So it’s fair to say that Silver is very influential when it comes to predicting elections.  And it’s no big surprise that he has more than 410,000 followers on Twitter

But consider this: if Nate Silver said that he brushes his teeth with CrestComplete toothpaste, how much influence would he have in convincing people to buy Crest? 

If Nate Silver speaks about CrestComplete toothpaste, don’t bet on many of us being influenced by his endorsement.

My point—and it’s the point that I think is so big its lost in the discussion of “scoring influence”—is that we follow popular people on social networks each for a special reason, and even with that the chances they will influence to do anything other than websurf some more is very, very small.

Nate Silver has a Klout score of 89.  Now, I’m no expert on Klout :-) but 89 is a big whopping Klout score—bigger than Thomas Friedman, the second most influential person at the New York Times apparently.

A 2010 study of Twitter influence by the HP Social Computing Lab (see pg. 8) showed that the most famous people on Twitter like Martha Stewart and David Gregory were very popular but not very influential.  That is, what they communicated had very little influence on their follower’s actions.  If Martha Stewart says “orange is the new black” nobody goes out and buys orange.

The foundational truth we’re not talking about in this discussion of digital influence is that popularity—even subject matter expert popularity—is not a proxy for influence.

Truly understanding how digital influence works is a challenge we’ve barely begun to understand. 

What does that mean for marketers?  Using content strategy, search optimization and social engagement are becoming more important for building relationships with customers.  But the solution for your individual company will be unique and will require you to find ways to build trust with your customers. 

And unfortunately you can’t influence your customers by cultivating something called “influentials.”