The second day of Predictive Analytics World was every bit as good as the first. Attendance was much greater than expected - yes, we are in the era of analytics!
The talk from Sun Microsystems was about reaching both customers who had more time than money as well as customers who had more money than time. John Wallace of Business Analytics and Traci Chu of Bella Pictures (a wedding photography small business) gave a very interesting talk. The photographer's goal was to improve her understanding of a bride's propensity to set an appointment to discuss wedding photography (and therefore likelihood of purchase). The marriage (Eric Siegel's pun-iness is catching) of modeling expertise and domain expertise was an effective combination to improve significantly over the previous approach. It was a real challenge for them to have a model in 6 days. John talked about "coding with an audience," addressing "data pathologies" and removing unnecessary variables more quickly as a team.
It was so good to finally meet Andreas Weigend in person. I had invited him to speak at M2003, but was unable to attend (maternity leave :o). His talk was engaging and informative with lots of interesting ideas to ponder further. He started by asking the audience what the bottlenecks were to doing analytics. The answers from the audience included:
- Data Integration.
- Mentality of some people.
- Data quality.
- IT.
- Communication.
- Documentation about what the data means.
I loved his concept of CRM moving to CMR (Customer-Managed Relationships). How nice that he concurred with a point from my presentation about the importance of experimentation. (I had included a quote from Derek Bok, former president of Harvard who had said "If you think education is expensive, try ignorance." How else can we learn if we don't try something new?)
The panel on Cross-Industry Analytics Challenges with Dean Abbott of Abbott Analytics, Richard Vlasimsky of Valen Technologies and my colleague, Vijay Desai was an interesting blend of experiences and perspectives - all well received. The recurring themes of data quality, measurement and deployment challenges were raised with lots of audience interest and discussion. I am looking forward to a nice dinner with some of the conference speakers before taking the red-eye home.
It was also great to finally meet James Taylor. I had to miss his talk Wednesday to meet with someone else, but heard it was also very well received. I thoroughly enjoyed the time here.
