How eBay plans to Capture Sales from Brick-and-Mortar Stores
The auction site hopes to expand its market by modeling consumer behavior.
Even though eBay is one of the giants of e-commerce, it’s limited by the fact that online sales make up only about 7 percent of all U.S. retail spending. Now it has its eye on the business that currently goes to brick-and-mortar stores—but it will need to rely on smart modeling methods to capture it.
First, eBay will need to know you’re about to buy something offline, so it can shoot a message to your smart phone advertising a better deal on its site. Making that possible is a task for 24-year-old Jack Abraham, who joined eBay this fall as its “director of local” after the company acquired his startup, Milo.com, for a reported $75 million.
Milo sits at the confluence of physical-world and online shopping: it’s an inventory search engine that allows you to see what is in stock at stores nearby. (They coöperate so they can see Milo’s data on what people are searching for.) Abraham wanted to bring the transparency of online retail to neighborhood stores, letting shoppers compare prices on the same products.
The idea behind Milo is that some purchases—say, TVs or clothing—are best made in person, yet product research typically begins online. Forrester estimates that 42 percent of sales at brick-and-mortar stores start with tools like this one, and that 53 percent—over $1.4 trillion worth—will do so by 2014, even if that purchase is just a bottle of shampoo or a cheap pair of jeans.
That’s where eBay sees a huge opportunity. When you search Milo, eBay will see that search as an indication that you might be open to a competing offer. As soon as you start clicking through Milo’s search results, eBay’s system can race through the vast number of products available on its site and try to divert you with a coupon, discount or related product. Even if it can’t, that’s all right, says Steve Yankovich, eBay’s vice president for mobile. “We want to be the starting point every time you’re thinking about making a purchase, because if you always start with us, you’re more likely to [buy] with us as well,” he says.
The company has other ways of knowing that you are warming up to a purchase. Its mobile apps, which have 14 million users, also feature RedLaser, a technology that lets you snap photos of product bar codes inside a store and call up more information on a product, including any better prices that might be found online. Like a search on Milo, a RedLaser scan tells eBay that a purchase could be imminent.
But eBay’s system still doesn’t know much about your motivation, so it’s going to need more data before it lobs an offer to your phone. At this point, Yankovich says, the system will dive into its enormous data sets. The company facilitates $60 billion worth of product sales a year and gets two billion unique product searches per month. Now, eBay can sift through that data to look for behavioral correlations that could help it guess why you are showing an interest in a particular brand or item.
For instance, eBay can look at your personal history on its site, trying to figure out whether that fishing pole you’re mulling is a one-off purchase—a gift, perhaps—or the kind of outdoorsy stuff you buy regularly.
Armed with those correlations, and knowledge of your location (if you allow that) and your Milo or RedLaser searches, the system can make some initial assumptions about the kind of offer you’re most likely to respond to. “We can see some totally surprising correlations with this data,” says Yankovich.
If eBay’s plan fails, it wont be the first time an ambitious data-mining project fell short. But Abraham says the vision is so compelling that eBay is likely to stick with it. “If we can create this unbelievable personalization engine,” he says, “we can show people offers that will be so compelling that they’ll change their behavior.”
Become an MIT Technology Review Insider for in-depth analysis and unparalleled perspective.Subscribe today