It seems every month there's either a conference or workshop from Predictive Analytics World. TDWI now has a heavy emphasis on PA with its quarterly shows. O'Reilly has made a big splash in the data science world with Strata. Gartner's BI Summit has a focus on predictive analytics. And these are just the ones I've participated in recently.
Despite the ample supply, I was intrigued when given the opportunity to attend the Business Analytics Innovation Summit from the IE Group in Chicago, last week. About 100 hearty souls from across the country struggled with bad weather and transportation delays to participate. Even to a jaded Chicagoan, discussing the wind chill index with out of towners less than a week before Memorial Day was embarrassing.
In contrast to its larger competition, BAIS was much smaller, with just a single track of presentations. I kind of liked that, not feeling the pressure of having to make choices from among equally attractive competing sessions.
Joe LaCugna started things off discussing Microsoft's perspective on a changing work world of IT consumerization, a diverse work force and cloud computing. His Delta analytics framework – Data, Enterprise, Leadership, Targets and Talent – borrows from Tom Davenport's “Analytics at Work.” LaCugna's discussion of the use of ethnographics to identify problems with customer satisfaction that leads to re-licensing “churn” was most interesting.
Jane Griffen of Deloitte Consulting is clearly an experienced pro at the use of intelligence and analytics to solve business problems. The principles she espoused that include starting with an assessment of current analytic maturity, identifying crunchy questions, looking for signal strength, accelerating insight to business processes and the need for organizational culture change, resonated with participants.
I liked her observations on the failures of many physician targeting campaigns associated with new drug promotion initiatives of pharmaceutical companies. Why did the campaigns fail? The targeted MDs were often long on reputation but short on referral networks.
Thumbs up to two presentations by analysts at Wyndham. Srinidhi Melkote and Jeremy TerBush discussed a multi-year effort to re-optimize Wyndham's vacation ownership exchange software. Their challenge is one of both forecasting and revenue/yield management. In the re-work, they let the data do a lot of the talking on the value of a February week at a desirable Hawaiian property versus the same week at a property in Branson, Missouri. The new system is much more granular and flexible than the old – resulting in more profit to the company.
Shankar Mishra VP of Online Retailing at Wyndham, spoke on maximizing analytic impact with minimal effort. I liked his analytic framework of linking business outcomes to drivers, which is akin to connecting leading and lagging indicators in the science of business. Outcomes include supply and demand, while drivers might be marketing spend, pricing attributes and staff development.
Mishra wants to know why outcomes end up as they do. To get the answers, as a company develops its strategy, it articulates a series of hypotheses that link cause and effect – drivers to outcomes. Intelligence and analytics are then in the service of measuring, evaluating and predicting the strength of these linkages. For Mishra, velocity of response is critical to the feedback loop of monitor, analyze, prescribe and act.
Among the presenters I heard, Deloite, IBM and Host Analytics were also conference sponsors. Sponsor/presenters have the difficult task of satisfying participants with content while at the same time promoting their brands – not always an easy task. I thought Deloite was the best of the three at this juggling act, followed by IBM and Host Analytics.
I, unfortunately, was only able to participate in the first of the two-day conference. So my grade for BAIS is a positive incomplete: I like the more intimate venue and was impressed overall with the presentations, but wasn't able to see enough to pass comprehensive judgment. I'll be able to submit a “final” grade after participating in the Predictive Analytics Innovation Summit in early November.