Database Marketing: Three Strategies for Successful Data Mining
Which data models are worthwhile? What are the best predictors? Which metrics work? A panel of catalogers and list pros provided simple tactics to help mailers improve the quality of their databases at the “Trick Out Your Data and Kick Up Your Revenue” session held during the List Vision conference earlier this month in New York. Following are their tips:
* Look for predictors within your customer data. For instance, Don Austin, director of client strategy for May Development Services, the nonprofit arm of list firm Direct Media, discussed a women’s apparel catalog for whom merchandise was a good predictor of lifetime value. Customers whose first purchase from the catalog was a knit top tended to have greater lifetime value than customers who purchased suits.
Somewhat less suprisingly, customers who initially purchased multiple items had a greater lifetime value than those who first purchased just one item. “Look back at your data to see which customers were worth more and see if there’s any commonality between their initial purchases,” Austin recommended.
* Determine which data models are worth it. Hosiery catalog and continuity marketer HCI Direct appended RFM data to its inactive customer file to reactivate older buyers. But the effort proved too costly, revealed senior director of database marketing Marijke Bekaert. The time, effort and money to append RFM data exceeded the benefit of the process. “You need to test whether it’s worth it for you,” she said. That said, she found that household age is often a good predictor of purchase behavior, so she regular appends this data to the 18 percent of her names that don’t have it.
* Establish metrics for success. “We often recommend using third-party data, data mining and data warehouses. But first, have the metrics [necessary] to evaluate testing and use of these analytic tools,” Austin said. If you don’t have the metrics, he said, then you won’t know if you’re doing well.
