DMNews talks with David Kimrey of

Q What was the marketing strategy before you installed the PASW Modeler from predictive analytics software provider SPSS?

A It was pretty much guesswork trying to figure out who was going to purchase what. Our e-mails were sent only to people who had specifically selected certain events or genres. We didn’t have any sophisticated forecasting or A/B testing for our site, so it was kind of like the Wild West.

Q Why did you choose to add SPSS PASW Modeler to your toolkit?  

A The idea was for me to come in and get the data warehouse up and usable so we could use more sophisticated modeling. We really needed the capability to do predictive modeling. When we made changes to the site or to e-mails, we needed to know whether their effect on consumers was positive or negative.

Q What was the end goal when you installed PASW?

A The business challenge was to increase sales, especially with the e-mail side. We didn’t feel we were targeting everyone we could have.

Q How did the Modeler affect your approach to marketing?

A One of the first things we did was build out customer clusters, and we came up with four distinct clusters of people. After the clusters were created, we started testing the messaging to different clusters to see how they would respond. We could then better segment and customize our e-mail messages.

Q How did that change response?

A The open rate on tested e-mails improved 24% over the baseline, while the click-through rate improved 58% from the baseline. The change made in content and targeted offers improved traffic in the overall e-mail channel and contributed revenue at a rate of nearly one and a half times the baseline in the targeted sends.

Q Moving forward, how will you keep improving your targeted marketing?

A We’re going to build more clustered profiles around each site to see the differences between people who come from Google, Yahoo or Bing and try to start changing the messaging of ads on the site to appeal to particular interests.

Related Posts