Q&A: IBM's Top SPSS Exec On Predictive Analytics

Marketing veteran Deepak Advani talks about business analytics, predictive models, and Big Blue's coming decision-management offerings.

...
As IBM's vice president of analytics, [Deepak Advani] effectively replaces Jack Noonan, the former CEO at SPSS, which was acquired last year by IBM for $1.2 billion. Advani discusses Big Blue's plans for one of the keystones of its business analytics and optimization practice.

InformatonWeek: How does SPSS fit into IBM's "Smarter Planet" strategy?

Advani: I don't know if it's the economic downturn, but a lot of companies are now looking exclusively at business outcomes. Yes, you need technology and speeds-and-feeds to deliver a business outcome, but companies are very focused on the end game. They will say, "We're a telco and we need to reduce customer churn. How do we do that?" Or "I'm a police chief and I need to do smart things to reduce crime in my area."

IBM's Smarter Planet strategy is all about taking the petabytes of new data that gets created every day and mining that for insights to build competitive advantage. SPSS fits in really well because it comes at that challenge from the angle of understanding social behavior. If we have a data-rich environment, we can build a model that can predict what the outcome is likely to be.

We can apply different statistical techniques to build multiple models and then find the one that gives us the greatest accuracy. With SPSS, we can go beyond the rear-view mirror and start predicting where things are headed. That's a capability that was missing from the Cognos portfolio, and I know Rob Ashe [formerly CEO of Cognos and now IBM's general manager of business analytics and Advani's boss] would agree with me.

InformatonWeek: Before it was acquired, SPSS was on a mission to broaden the appeal and use of its tools beyond PhD statisticians and data-modeling types. Will that continue under IBM?

Advani: SPSS really did grow up out of the academic community, and the tools really appealed to mathematicians and statisticians. Once they got their degrees and went off into the commercial world, they stuck with the tools with which they were familiar.

Over the last couple of years, SPSS had focused heavily on ease of use. If you look at our modeler, for example, the collaboration and deployment tools are really targeted at business users and business analysts. The interfaces are very graphical, so business analysts can build models without having a heavy dependence on mathematicians, statisticians and programmers.

Read more.