'Really, I can return it? Sold!'

Eric Anderson models customer data to understand exactly how much individual customers will pay for a liberal return policy

By DeLene Beelandan

	Eric Anderson

Around Christmas, consumer Celeste Pramberger visited a Babies "R" Us in Florida to exchange an outfit for her baby, Allie. But the store did not have Allie's size, so Pramberger asked to return it. She was mortified when the clerk said she could not because she had returned too many past purchases.

"I was so surprised," Pramberger said. "The only things I had returned had been four months prior from my baby shower registry."

Although her returns ban occurred five years ago, Pramberger said the incident left her reluctant to return anything for fear of being embarrassed. And it may have left her gun-shy to purchase items in the first place, for fear of needing to return them later.

People return items for a multitude of reasons, from poor fit to buyer's remorse. And some stores, like Babies "R" Us, track their customers' behavior over time and then tailor their return policy based upon the customer's personal shopping history.

But are they harming or helping their profit potential by doing so? A new model developed by a team led by Kellogg Professor Eric Anderson can help retailers determine just that.

"Every retailer has a return policy, but there is no empirical evidence as to how effective they are or how much customers value these policies," says Anderson, the Hartmarx Research Professor of Marketing. "There is just not much evidence about returns in general, although this is an emerging field in marketing science."

The model can help retailers optimize their return policies for individual customers as well as across product categories and purchasing modes — for example, online versus in stores. And because the model monetizes the return option, Anderson says it can also shed light on gross and net demand changes based on the option to return, and calculate more accurate price elasticities.

To pinpoint the numbers, the team — which included former Kellogg professor Karsten Hansen, now at the University of California's Rady School of Management, and Duncan Simester, a professor at MIT's Sloan School of Management — devised a statistical model that analyzes individual purchase histories and predicts customers' purchase and return rates within defined product categories over time.

The model also calculates how much customers would pay for the option to return purchases within these categories and how a retailer's return policy affects whether or not they buy a product.

For example, Anderson's team used household data from 987 individuals over 10.5 years and analyzed their purchases for women's footwear, men's tops, and women's tops from a single retailer that sells through stores, a Web site and mail catalogs. They calculated that while men were only willing to pay $3.19 for the option to return tops, women were willing to pay $5.

But the real surprise lay in footwear. When the average price for shoes was around $50, women were willing to pay an extra $15.81 per pair if they had the option to return them later.

"One way to think about that is if you were a direct marketer and people didn't have the opportunity to return shoes that they hadn't tried on, you'd have to drop the price down to $35 to sell those shoes," Anderson says. "For retailers, that is a big deal."

Anderson's research provides the type of information a retailer could use to customize its return policy to individual customers. And that's not all: It can also help retailers measure how a change in the return policy will shift both gross and net demand, and then calculate the expected changes in net profit. Retailers can also plug their own customer data into the model to optimize their return policies across product categories and across selling modes. Finally, the model can help retailers calculate more accurate price elasticities by incorporating changes to their net demand when returns are allowed.

"This model is quite general and can be applied to many types of products, not just retail clothing," Anderson says. "The model is applicable in any situation where the customer has the option to renege on a purchase."