Golfsmith lowers costs with predictive response model

Client: Golfsmith?

Vendor: SAS Analytics?

Objective: Reduce cost and increase incremental sales while segmenting customers according to purchase behavior. ?

Golfsmith International, an Austin, Texas-based specialty retailer of golf equipment, sought to improve direct mail response rates, data merging costs and campaign response times. With 75 retail stores, an e-commerce site and a direct mail catalog, the company needed to track data across channels to better understand customer behavior and build more targeted, effective campaigns. ?

“The company was hungry for data,” says Mu Hu, director of CRM at Golfsmith. “We didn’t know our customers. We had a third-party data provider that did data hygiene and segmentation, but we needed to increase incremental sales and reduce costs.” ?

STRATEGY: Golfsmith chose SAS Analytics in 2009 to help build a predictive response model and perform data merges. Hu says he chose the Cary, N.C.-based vendor because he had worked with it during his previous role as manager of database marketing at Zale Corp. SAS, he says, provided Golfsmith a fast, flexible tool that “could do anything we wanted it to.”?

With SAS up and running, the retailer began to segment its customers into several buckets: Live for Golf, Avid, One-Time Shoppers, Reactivation, Inactive Best, Inactive Average and Big Loss. Golfsmith sent different offers with different creative messages to each segment. For example, Live for Golf customers were sent cross-sell and up-sell messages as opposed to Inactive Best customers, who were sent messages informing them they were missed and offering 10% to 15% discounts. ?

“We performed tactical level segmentation that [revolves around] purchasing behavior,” says Hu. “The idea is to ?segment those customers by channel to identify which vehicle is best for each customer.”?

Customers were also segmented into buckets according to preferred method of communication: All Channel, Phone and Retail, Phone and Web, Phone, Web and Retail. ?

RESULTS: By the time Golfsmith put together the June 2011 mailer, it had improved merging procedures and reduced costs by 50%. The company had identified consumers most likely to respond to the mailer and sent out 20% fewer pieces of mail than in 2009. Golfsmith did not send out a 2010 June mailer. ?

With improved segmentation and fewer pieces mailed, the June campaign generated 30% more responses and an incremental sales increase of $500,000 compared with the same campaign in 2009. ?

“Before SAS, every campaign we mailed to the same customers again and again,” says Hu. “With SAS, I can use a response model to select customers who are most likely to ?respond to that particular campaign.”

Related Posts