How much will each customer spend? Startup Zodiac says it has the answer
March 24, 2016 02:12 PM
Thoryn Stephens has long believed he could market more effectively if he could calculate the worth of each customer. That’s why Stephens, chief digital officer at American Apparel Inc., jumped at the chance late last year to be among the beta testers of new technology from Zodiac designed to predict the lifetime value of each of a retailer’s customers.
Zodiac announced Wednesday it has raised $3 million in seed funding to further develop its product, which is based on decades of work by Peter Fader, a professor of the Wharton School at the University of Pennsylvania, on how to calculate customer lifetime value. Fader’s connection helped sell Stephens on trying the system. “He is the king of customer-centricity,” Stephens says.
Stephens sees several benefits in knowing how much a customer is likely to spend. First, a retailer can better segment its best customers and make sure they get offers that keep them engaged. It can also find consumers like its best customers, for example using the Facebook Custom Audience ad system, and target them with ads.
A retailer also could analyze the behavior of its high-value customers to derive insights that could induce mid- and low-value customers to buy more. “Maybe recency is a key factor,” Stephens says. “What if you sent each customer an email within a day after they visited the site or bought, would that drive additional revenue?”
American Apparel, No 330 in the Internet Retailer 2015 Top 500, has provided Zodiac with a portion of its customer file and obtained useful insights into how customers break down in terms of lifetime value, which Stephens says is important for understanding “the health of our customer file.” It’s also helped the clothing manufacturer and retailer begin to target its best customers, although Stephens says it’s too soon to gauge results.
A big question for any system that attempts to predict customer behavior is whether it can do so accurately. Stephens says American Apparel provided Zodiac with a portion of its customer data and received back results that were “in line with the internal analysis we’ve done.”
Fader, the Wharton School professor, has been working for nearly 20 years to perfect statistical models for predicting customer lifetime value, and collaborated with three of his students to turn it into the commercial product that Zodiac is now selling, says Scott Samios, chief operating officer of Zodiac. He says the company launched early in 2015 and in recent months has recruited a half-dozen companies to test its system. Besides American Apparel, they include Dressbarn, part of Ascena Retail Group, No. 87 in the Top 500; ModCloth, No. 198; and cosmetics brand Charlotte Tilbury.
Samios says Zodiac proves its accuracy by taking historical customer data from a client company, for example the past two years of transactions, and analyzing the first 18 months of that file to predict how much customers would buy in the last six months, without looking at that last half-year of data. He says in most cases Zodiac’s estimates are accurate within 3%.
He says a key benefit of Zodiac is that it predicts the spending for each customer, rather than relying on averages for a group of customers, which inevitably overstates the value of some customers while undervaluing others. In addition, rather than relying on historical data such as average order value, Zodiac updates its estimates based on current transactions, sending an updated estimate of each customer’s projected value each month.
Samios says Zodiac also answers the question: What is a customer’s “lifetime” with each retailer? Most retailers pick an arbitrary period, such as one or two years, Samios says, while in many cases a shopper may patronize a retailer for a much longer or shorter period. Zodiac, he says, can reveal what the typical time period is for each client retailer, he says.
Zodiac charges a monthly subscription fee that starts under $1,000 and goes up based on the size of the customer file a retailer wants tracked and analyzed. “Even a large brand might start with a good-sized cohort and be able to get started for around $1,000,” Samios says.
Stephens, who joined American Apparel just over a year ago, says American Apparel is seeking to make better use of data to drive its marketing, but doesn’t have the data scientists on staff to optimize results from the marketing technology it’s using. “Outsourcing made the most sense, given the internal infrastructure I’ve inherited,” he says. “I knew I could get to market quite quickly working with the Zodiac team, which is exactly what’s happened.”
Leading the $3 million funding round was First Round Capital, a venture capital firm that specializes in early-stage companies, Metamorphic Ventures and others, Zodiac says.