Our series on choosing the right analytics for the job continues with regression models. Regression models enable businesses to predict how individual customers are likely to behave. With such specific insights, businesses can differentiate between customers to a much greater degree, further increasing the granularity of segmentation and the relevancy of offers. In fact, businesses can go as far as to essentially create segments of one.
Regression analysis delivers this level of specificity and accuracy because while it can encompass a vast range of internal and external data, its power comes from pinpointing the specific customer attributes most predictive of a future behavior.
Relationships between numerous attributes and other variables are examined to see how a change in the value of one variable affects the value of another, dependent variable. Attributes that prove to be highly predictive of a behavioral outcome are incorporated into a predictive model. For example, a regression model can be built to predict a customer’s propensity to make a purchase in a product category or to discontinue using a service.
Because these models can predict for each individual customer the likelihood of such an outcome, they open up the possibility of unique treatment. Moreover, by using multiple regression models, retailers can gain a much clearer picture of the customer. Knowing that Jane is not only likely to buy a 48 inch TV, but that she tends to like the Sony brand, but doesn’t tend to go for cutting-edge products, enables the retailer to greatly increase the relevancy of individualized offers and interactions.
Here are a couple examples of how our clients have applied such insights:
- By implementing propensity models for all of its product categories, one client could dynamically generate individually tailored emails for customers. Every customer was scored for propensity to buy in each product category. Based on where the customer scored the highest, the client’s automated decision system determined which of several overall mailing themes was most relevant, then added features or offers in six product categories. No more than 20 customers in a million received the same set of recommendations. This kind of tailored promotion enabled the retailer to achieve email open rates of up to 50 percent and up to 9 times the ROI compared to generalized mailings.
- Another client used predictive analytics to build membership and activity in its fee-based premium loyalty program. The program delivered 12 individually selected relevant offers per month to each member; these were loaded onto the customer’s loyalty club card and could also be accessed from the retailer’s website or from kiosks in stores. Each offer package included rewards aimed at increasing buying frequency and customer retention. It also included an offer encouraging the customer to try a product in a relevant category where they had not purchased before. In the first six months alone, this client increased program membership by over 43 percent.
Tune in next week when we’ll address time-to-event models. If you can’t wait for the next post on this topic, check out our Insights white paper: "Which Retail Analytics Do You Need?" (registration required).
