Skip to Content

EBay: The OS for E-Commerce?

The online auction giant wants developers and niche marketers to tie into its infrastructure, says eBay Research Labs’ senior director, Eric Billingsley.
August 9, 2006

Microsoft has had a research division since the early 1990s. Google’s programming staff spends 20 percent of its time on R&D. Even Yahoo has a growing research division (see “Yahoo Ramps Up Research”). So why not eBay?

In fact, the online auction giant has a research outfit, although it’s little more than a year old. Known as eBay Research Labs, it began as a corps of programmers called the Advanced Technology Group.

For its tender age, the division has outsized ambitions. Its senior director, Eric Billingsley, a former nuclear engineer and veteran of search industry pioneer AltaVista, believes eBay will become a “common platform” for all kinds of online commerce, allowing people to build on and profit from customized tools that tap into eBay’s massive inventory and customer-service infrastructure.

By cultivating new ideas inside the company and also reaching out to third-party developers, Billingsley is trying to position eBay as the New York City of the Internet economy – the place every independent developer or startup CEO with e-commerce ambitions has to go to make it big.

Meanwhile, the lab’s researchers are analyzing eBay from within and without – improving the backend hardware and software that keep the existing online marketplace running, while coming up with new ways to make users feel more at home when selling and buying on the site.

Technology Review senior editor Wade Roush interviewed Billingsley at eBay’s headquarters in San Jose on August 2.

Technology Review: What was your first big project within the labs?

Eric Billingsley: The first one we really dug into, as the Advanced Technology Group, was the search engine for eBay. Back then, it would take nine hours to update eBay’s index. Searches were extremely slow, and it was becoming a very expensive part of the infrastructure. Now, when you place a bid it’s a matter of seconds before it shows up in the index. So right off the bat, we started having some impact. And that got us some clout and made it possible for us to start doing some other things within eBay.

TR: How many people do you have in the labs now, and what types?

EB: I’m a nuclear engineer by training. I’ve got two physicists, I’ve got an anthropologist, I’ve got a number of computer science people of course. Mathematicians, statisticians, computational linguistics, machine learning. Right now it’s a very small team.

TR: It sounds like you’re not a classic research division, where there are lots of PhDs off in a building thinking of cool ideas.

EB: That’s what makes it an exciting job for my guys. I don’t want to hire people who want to be in an ivory tower. We want people who are going to invent the technology that’s going to power everybody’s infrastructure in five years.

With that in mind, one thing we’re looking at is how we can apply some ideas from social networking. We’re taking the feedback of our users and using that to actually help the community find what they find most interesting. Have you looked at eBay Express yet? That’s the first iteration of some of the finding work that we’re going to be putting out over the next year.

Normally at eBay you’ll find quite a mixture of things. Let’s enter the search term “iPod nano.” [Billingsley turns to his laptop.] You get earphones, chargers, covers, armbands – pretty much anything except an iPod. But at eBay Express we know [from studying previous users’ behaviors] you either want iPod accessories or you don’t, so those are the choices you get. That allows you to very rapidly zero in on what you’re interested in.

TR: So you’re giving people a solution for the frustration of getting ten different kinds of things when they’re really just looking for an iPod. Are there other frustrations or inefficiencies that you’re trying to help people solve?

EB: We’re trying to get better at understanding the intent of a buyer. The way we see it, if you’re looking for a piece of electronics, you are thinking about very different things than if you’re looking for a glass collectible. You’re thinking about, how does it fit into my living room? Does it take U.S. power? What’s the shipping cost going to be on this thing? You’re going to get a list of products and you’re going to make a comparison between this, that, and the other.

But when you are searching for a collectible, you’re thinking about, what do I already own, and how do I find more like that or something completely different? What’s of value to you may be very different from what’s of value to someone else. We’re trying to get a little further down the road of making those into customized experiences.

TR: Okay. I own an old wooden stereoscope, from 1905. You can buy cards for them that have two photographs taken from slightly different angles. It sounds like you want to make it so that I can click on a few buttons and get a special “Stereoscope Store” to find more cards.

EB: Exactly. But my real goal is to create a platform that allows third-party developers to develop that store for us and that allows them to profit from it. I want to build eBay such that it’s almost a plug-in architecture, and we can allow developers to go in and create these perfect buying experiences within these little niche markets. Because that “long tail” is huge. And we’re only beginning to tap it.

TR: Say more about this plug-in architecture and how it would benefit outside developers.

EB: This is still in early phases, and I’m talking about it openly because I’m looking for the rest of the Internet community to help us with this. I want to enable people to build a business on top of a business on top of a business. I want to bring our infrastructure stack to the entire developer community such that they are no longer limited by having to buy and run their own infrastructure. Ideally, we want to take it to the point where a single developer anywhere in the world can make a living developing applications for a company website and never actually be an employee of that company.

TR: You have some peers not too far away from this campus that are doing similar things. With S3, Amazon is encouraging people to basically offload all their database operations to Amazon’s excess capacity (see “The Internet Is Your Next Hard Drive”). And you’ve got Google persuading people to upload their entire portfolio or inventory into Google Base.

EB: But all of those are very limited services. It’s all about, “Give us your data.” And from research labs, I can tell you, data is power. So of course it makes sense for companies to do that. But what I want to do is bring the traffic to the developer. I don’t expect anybody to do anything for free. When you list something with eBay, it’s because you want to make money off that transaction. We have people who make their livings doing that. What I want to do is expand the types of people who can do that. I want not only people who have an inventory of products, but also people who have an inventory of ideas, education, and ability.

TR: And these new applications you want these people to build – do they have to have something to do with selling?

EB: I think everybody has their purpose in life, and eBay’s purpose is to pioneer new communities built on commerce. Commerce is behind almost every aspect of modern life. I don’t see how that limits us in any way.

TR: You said that you guys brought yourselves into existence by proving that you could rework the search mechanism on the site. I’m wondering whether at the same time there might have been a sense that eBay was reaching a certain size or a certain revenue level where it really ought to have an R&D operation.

EB: I’m sure you’ve heard about the “train seat” model at eBay. The train model is that we completely re-release the entire eBay site every two weeks. We just launched Train 472 on Monday, so everybody who had seats on that train saw their project go live on the site this week. That gives us predictability – we know [when trains will arrive] a year ahead of time. The problem then becomes that our engineers are booked a year out, and there’s not much wiggle room. So we’ve always had in the back of our minds that you need to have an extra set of resources – people who are thinking about not just the business ideas, but how we can take new technologies and create new ideas that we can then feed back to the business.

TR: So the folks in the research lab aren’t assigned to train seats.

EB: No.

TR: With all of the new features you’re building, you’re making it possible for people to spend a lot more time on eBay.

EB: Well, I’m hoping that by spending more time on eBay, they’ll have a better life as a result.

TR: Or they’ll have more stuff. One or the other.

EB: Some people feel complete once they’ve got that complete collection. I want to make them feel complete.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

How scientists traced a mysterious covid case back to six toilets

When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.