Chaotic influence or why my cat could be a VIP

There's been quite some noise when Gladwell published "Tipping Point" and it's a well written book with an interesting theory about Salesmen, Mavens and Connectors. But does the theory really work? Are there always Salesmen, Mavens and Connectors in all fields of human activity? Can the process be "artificially" triggered and guided to help someone deliver a targeted message to a given audience or achieve a specific kind of behavior in a population (e.g. buy something)?

Considering the vast and increasing sums of money invested by marketeers these past few years one would think that yes is the answer to all of the above questions. Maybe not quite... no make that "definitely not". So argues Duncan Watts who challenges the very idea that influence can be "remote controlled" by creative types working for ad agencies. He makes the case that influence is in essence a chaotic process such that anyone could in fact and perhaps unbeknown to themselves, exert influence. He goes even further to claim that influence happens by accident more often than not. So according to Watts, my cat could very well be a VIP, i.e. a Very Influential Person (yes, my cat is a Person)! The article of Fast Company is worth reading. Interestingly, Watts speaks of a chaotic process and argues that
influence happens mainly by accident (see graph below scanned for Fast Company). he seems to have reached that
conclusion by programming artificial populations with hundreds of
different combinations of parameters in the models of "influence" or
transmission he used. There was no conclusive evidence as to the existence of influential individuals who are supposed to be able to trigger the spread of a message of behavior across a large proportion of a population.

Now, the very concept of "influentials" is something that drives me nuts because there is a hell of a lot of talk about them in all sorts of campaigns, yet nobody seems capable of giving a rigorous definition of what that new animal is, nor how to characterize and measure the phenomenon. Bottom line: there is considerable lack of precision and quite some BS in the stuff being told to people attending seminars and to advertisers who are so desperate to actually make an impact that they are willing to trade the unmeasurable means they know for supposedly better means that will deliver better ROI. And in fact the early adopters have been quite successful, but as the number of "influence marketing" operations increases, the impact of each of those campaigns will become lower and lower. So unless online marketing professionals actually tackle the imprecision regarding the very concept of "influence" I have hard time seeing this business become sustainable in the long run.

If I were in that business, I would invest some research money into building a minimalistic model of the influence phenomenon based on a model coming from the field of complex adaptive systems and I would relentlessly test and refine it to reach a point where entire campaigns could be structured in a more scientific way with the end objective to be able to identify fields in which campaigns can actually rely on "influence" and if so to make those campaigns as efficient as possible.