AAA Auto Club South’s Ernesto Diaz and Kristin Rahn on Efficiently Building Predictive Models
AAA Auto Club South has always been an expert on helping its members efficiently get from point A to point B. So when the Tampa, Fla.-based affiliate of AAA had a chance to improve its own efficiency while building predictive models, the association locked in the opportunity.
In June 2009, AAA Auto Club South hired Britain-based integrated marketing solutions provider Portrait Software and began using Portrait Customer Analytics. By August, the affiliate had 6.7 million household and customer records added into the system.
Now, says Kristin Rahn, the affiliate's advanced analytics director, building predictive models takes less than a third of the time it once did. Plus, says Ernesto Díaz, the affiliate's managing director of insights, the association's staff doesn't need computer programming backgrounds to manage its modeling activities.
Díaz and Rahn elaborate on how the affiliate is improving its predictive modeling efficiency and what comes next.
Target Marketing: AAA Auto Club South built predictive models to increase renewals, reduce churn, and cross-sell and upsell its members. How did analytics software aid in that goal?
Ernesto Díaz: I think it would be reasonable to say that from a productivity perspective, the building of those models—the production curve, if you will—was pretty quick. So we're moving into the next phase now, to test against those attrition models some executions to see if we can impact the renewal rate with 10 percent different tiers of our member experience. So we have members that have been members for maybe just a couple of years, and the attrition curve there will be a little bit different than members who have been members for between, say, four and six years and so on. ... And then we've also done a little bit of work on just trying to recreate a segmentation schema that we built about a year or so ago. ...
TM: What data did AAA Auto Club South input to build these predictive models?
ED: We are using all of our active members' membership information. We are adding to that demographic information that we have available from our demographic data providers. We're also taking behavior information—so purchases, active policies and those kinds of things that they may have with us. ... There's another aspect of it that we have that I would probably subcategorize under behavior information, like things that they purchased that we call our "relationship index," which is really measuring engagement of members with different product lines.
Kristin Rahn: Well, there's two pieces of it. And the piece that my team is using to do the statistical modeling and the profiling and targeting is fully implemented. ... And then there's another piece that's going to give a quick count and self-serve analytics to the business, and we're just rolling that piece out now.
Right now, we have two main tables that we've pulled together. One is at the household level. And we have 2.5 million records in it. And then we have one at the individual, or customer, level and there's about 4.2 million records in that. And we probably have 200 different fields in each of them. ... We have all kinds of behavioral and purchasing data at the household and customer level. And then we have a lot of demographic data that's external data that we purchase and append. And then we also have all of our marketing preference, opt-in or opt-out, for different business lines ... And with those, we can do pretty much anything we want to do without having to build new data all the time.
TM: How is AAA Auto Club South building segments, then determining how to use them in predictive models?
ED: The aggregate of the deciles would be ... the segments. ... So we're using the top deciles as the segment we want to target. In the case of the attrition model, we're trying to use the middle deciles as the segment that we're trying to pursue. Because the higher deciles would be people that are most likely to renew, anyway. So we want to focus our attention on the ones where we might be at risk of not renewing.
TM: Where is AAA Auto Club South gaining efficiency in building predictive models?
KR: Because the user interface tool is so easy, that's where we save time in building the models. But once we have the model, it lets us export code to score everything in batch. So all the scoring can be done in jobs overnight. ...
ED: Right now ... our focus is on increasing productivity of building the models. And, like Kristin said, having the tool is very user-friendly and efficient in doing that so that we can batch-feed this information at night and still get a significant improvement. Whether it's at call centers or whether it's just getting in front of who the customer is at the call center or the branch or doing direct mail and, ultimately, when we're working online.
TM: What advice can AAA Auto Club South provide to marketers who would like to build similar predictive models?
ED: One of the things that we focused on on the membership attrition side was ... the relationships that we had with the members at the time that the model was run. ... What we try to understand is at the time that we are making the decision to send direct mail, that's the time where we go back. And one of the things that we go back to understand is, first of all, the demographics. So to what extent the demographics play a role in a customer's propensity to renew or not renew. The other thing is to what extent does the relationship play a role to renew or not renew. And what I mean by the relationship is: the types of products that they have used, the number of times they have used the product and so on. So … that kind of data was the data we generated to provide to the modeling engine to then determine what are the variables that are most likely to help us understand whether a person will renew or not.