“CLV gives me a different view from just measuring reactivated or new customers,” said Michels. “It helps us move away from overall numbers in favor of quality.”
Dress Barn uses the product to segment its inactive customer base. Millions of people have shopped at Dress Barn but aren’t necessarily Dress Barn shoppers. It wouldn't be feasible to try and reactivate all of those customers, Michels said, so “[w]e weed it down based on quality and potential value.”
There are also big differences between shoppers coming in through different affiliate links. Someone coming through a coupon distribution service like RetailMeNot “is there for the discount, and doesn’t have a long-term stickiness,” Michels said.
By running all sales data through Zodiac’s CLV engine, and then feeding the data back into its CRM platform, Dress Barn segments out those with low lifetime value according to the algorithm.
The process doesn't downgrade all affiliate shoppers. Michels said shoppers referred by a company like Upromise, a loyalty program that accrues savings to college loan plans, “look a lot more like our regular shoppers,” so there’s a chance to turn some of those folks into Dress Barn customers.
Thoryn Stephens, chief digital officer at American Apparel, said he was referred to Zodiac’s CLV product at a Wharton alumni event, and similarly noted its benefits in separating the retail wheat from the chaff.
"We were looking for a way to segment out negative acquisitions, like serial shop-and-returners,” said Stephens.
To test the product, Zodiac COO Scott Samios said the company will regularly analyze a retailer’s previous two or three years of CRM data. “And we’ll tell them, ‘Hold the last six months behind our back, and we look to have results within 3% plus/minus accuracy.’”
And since the data is taken from CRM and goes directly back in (i.e., is associated with PII), Zodiac can be used to support digital advertising efforts using audience onboarding and matching systems such as those offered by Facebook, Google, Acxiom and others.
For example, Stephens said he can export an American Apparel list to Facebook for advanced ad targeting or can work with demand-side platform partner MediaMath to build lookalike audiences in other channels.