Predictive modeling is not a new tool for the retail business, eBay CEO Devin Wenig says, but in the last year it’s gotten good.
Really, really good.
“Something changed in the last year,” Wenig said Thursday at the Code Conference in Rancho Palos Verdes, Calif. “I think it’s a combination of computing power—GPUs being brought online to [manage] big data sets—and the sophistication of algorithms.”
Which is why eBay (EBAY) is all in. Fresh from its split with payments company PayPal, the e-commerce company wants to overhaul how it targets and serves its customers.
“It’s all about consumer relevance,” Wenig said. “I want it to feel like my eBay. You know [that] I want a $7 scarf or a $70 scarf. You know whether I like blue or red…I actually think we’re getting pretty close to that.”
Today, a registered eBay user will see curation on the eBay.com homepage. “What you see won’t be the same thing as 160 million other people,” Wenig said. “And if you start shopping and browsing you’ll start to see recommendations that I would argue are pretty good.”
Get Data Sheet, Fortune’s technology newsletter.
But it has bigger plans, which will begin to materialize later this year.
“In the last year we’ve gotten really serious,” Wenig said. “I would argue it’s more important for us than our competitors because of the breadth of our inventory. That’s why we’re marching down building the world’s biggest catalog—because we need to have visibility into that inventory. Search is a tool that has marginal diminishing returns when you have that large of an inventory.”
The company has challenges. Despite its successes, such as its long-ago acquisition of popular ticketing marketplace StubHub, eBay remains lodged in the mind of many users as a place for novelties, not critical goods. In short: It has a brand problem, even with almost $9 billion in annual revenues.
“[In the early days] it was cute to receive a roll of toilet paper from someone in Montana,” Wenig said. “We sell $90 billion of stuff now—and it’s not tchotchkes and toilet paper.”
The commerce market is worth an astounding $17 trillion, Wenig said. “I’d love the field to be dominated by three big players, but it’s not going to be that way. There will be lots.”
But only three have the kind of global scale to matter: Alibaba-owned Taobao, Amazon, and eBay.
“I don’t think there’s going to be [only] one company in the world where people shop,” he said with a tinge of frustration that eBay isn’t recognized as a commerce leader.
People shouldn’t be distracted by the hot new e-commerce site, Wenig seemed to say. (A recent ascendant example: Jet.) Real sales matter.
“There are lots of players with extraordinary growth that don’t sell in a year what we sell in a day,” Wenig said.
Plus, at least one large eBay rival sells things below what they’re worth, he added, a clear shot at Amazon.
“I don’t think that’s a business,” Wenig said. “I can sell a lot of $1,000 TVs for $500. That’s not a business.”
So what’s in store for eBay? Going beyond the World Wide Web to sell goods, Wenig said.
“I think of eBay as a commerce fabric” for people’s inventory, he said. So much of that activity won’t be on eBay.com. “I want to see eBay be ubiquitous where people spend their time,” he said—in other words, a platform. “An Internet that comes to you rather than an Internet that you go to,” he added.
Key to that effort is artificial intelligence and machine learning technologies and a broader harnessing of waves of user data.
“If you don’t express intent and I’m in your face trying to do something, that’s incredibly annoying,” Wenig said. “But if I express intent I’d love it to come to me rather than me come to you.”
Or, put another way: “E-commerce is a data business. We’re a data company. That’s what we do.”