Profiling, Segmenting Helps Jos. A Bank Move Beyond RFM

Fashion apparently knows no economic slowdown. Jos. A. Bank Clothiers has been in expansion mode since 2000, when it maintained 100 stores. At last count, it had passed the 500-store mark, with plans calling for further growth.

The company isn’t pulling back on its direct marketing plans, either. According to John Lewis, director of database marketing, the apparel firm’s challenge is not only finding new customers, but investigating more efficient ways of keeping the customers it has.

Jos. A. Bank had been using RFM (recency, frequency and monetary) analysis to segment its customer file, pulling campaign targets based on their transactions. As part of a test, Lewis sent transaction information from multi-channel buyers to data services firm Experian, which overlaid the records with compiled data and built propensity-to-respond models.

Jos. A. Bank’s database houses more than 4.6 million individuals who have made purchases or information inquiries via any of its channels during the last four years—3.2 million of those have done so during the last 24 months. Lewis did not indicate how many of these were multibuyers, nor how many Experian analyzed.

The next-purchase modeling provided a different ranking of prospects than Jos. A. Bank would have gotten from RFM segmentation. While Lewis didn’t volunteer campaign quantities, he did say that the overall effort levels remained the same. But response rates jumped 19.5%, and sales rose 16.7%. (Overall, the company distributed 9.6 million catalogs during its most recent fiscal year, up from 9.2 million in fiscal 2009, according to its most recent annual financial statement. It did not disclose the number of emails it sent.)

Jos. A. Bank’s next challenge was to convert one-time purchasers to multibuyers. “One-time buyers have so little transactional data,” Lewis says. “It’s been difficult.”

But not impossible. Again, Experian appended compiled data to these customers. This time, it used the data to create six customer personas, based largely on the type of clothing they had purchased and a variety of demographic and psychographic information.

“What we did with that was adjust mailing quantities aimed at each of the personas,” Lewis says. “It was a bit of a guessing game, when we were looking at what type of persona was going to get what type of promotion. We went deeper on ones most appropriate for [a specific] type of promotion, and pulled back otherwise.”

The company tested the Experian predictive modeling method for five campaigns before gaining confidence it would consistently trump its previous pre-campaign activities. It has, according Lewis, abandoned its RFM analysis in favor of the Experian method. In its May campaign, Jos. A. Bank sent out two direct mail solicitations: One for individuals who had bought suits, which featured a tailored gray suit and conservative tie, and another which highlighted a more casual tweed camel-colored sport coat.

The next step, for Jos. A. Bank, is to apply both the models and the learning to e-mail and other media, according to Lewis. While the company publishes its commercials via social channels “We have been told that we better be in social media, but we haven’t figured out how to make it work,” Lewis says.

Lewis shared his experiences during a session at the Direct Marketing Association’s All For One marketing conference.