Skill sets in predictive modeling

MM: As web analytics goes from more or less forensic reporting of anonymous behavior to now predictive models of what customers are likely to respond to, the underlying analytic discipline shifts from quantitative methods to things that are more forward-looking—be they game theory and predictive modeling.

When we talk about, “Who has skill sets in predictive modeling and econometrics?” Aside from analysts that are working at hedge funds and in New York City, and getting paid $300,000 to 500,000 a year, it seems to me that agencies will step forward into that opportunity with a founder or some equity participant. Secondly, it seems to me that there’d be a natural gravitation toward the largest concentration of PhDs in mathematics—India. And India may emerge, in fact, as the center of excellence for more prescriptive analytics and modeling of games, promotions and brand interaction through a lifecycle.

BK: I guess in response I’d say that the challenge the web analytics vendors play or have right now is that by and large, customers—people using web analytics tools—are implementing 20 to 25% of the capability of the solution to begin with. Limiting their use to data collection, analysis and reporting.

So, the more advanced, predictive modeling that you’re speaking to … while it might be something that they’ll ultimately build towards, the question is, “Will a majority of companies take on the task of implementing this capability?”

MM: I think that the agency will make that an opportunity and will invest in the tools and technologies to automate the tagging of your pages, and the reporting that comes from that, as well as the interpretation of what that means with very specific prescriptive outcomes in terms of what we need to build next.”

BK: I totally agree with you. I think it’s going to have to be a service provider. It’s not going to be the end company that’s going to try to have that expertise in-house.

MM: Yes.

Posted in: