“There’s been a lot of advances in machine learningthat take things that would have been literally impossible ten years ago and made those things much more possible today,” says Etsy CEO Josh Silverman. “With 62 million products for sale, picking for any given buyer the 20 or 30 that should be on page one of search results is a pretty interesting and pretty challenging task. The key is understanding what an item is with relatively little data and then being able to determine for each individual person how to personalize search results.”

Josh Silverman, CEO of Etsy, discusses how Etsy has increased growth by standing out in a world of sameness and by employing machine learning technology to personalize the Etsy experience for their customers. Silverman talks about his strategy for success in an interview with Fortune:

We Started Doing Much Fewer Things Much Better

Etsy has never been more relevant. In a world where so many of our products are being commoditized and we’re surrounded by a sea of sameness, Etsy stands for something really different. I think it’s really important that we stand out in the world and I’m proud of what the team has done to achieve that. The definition of success was really clear. I think from day one it’s about growing the size of the pie for everyone. The actual tactics that it was going to take to do that we’ve learned together as a team over time. 

When I arrived, there were maybe eight or ten different metrics of success that we all held relatively equally. I said there’s one metric that matters much more than every other, which is what we call gross merchandise sales. In other words, the total sales of our sellers. When we stopped saying what’s a good idea, what moves any one of these 10 metrics and started saying, what are the fewest things we need to do to really accelerate gross merchandise sales, we came to a very different answer. We started doing much fewer things much better. That’s really been the key to our success.

There’s Been a Lot of Advances In Machine Learning

Change is hard. When running a marketplace we have access to a lot of data and insights that each individual seller won’t necessarily have. Our job is to really look after the good of the whole and be willing to make some decisions that sometimes, in the moment, may not feel obvious to every seller but really do lift all boats and make our sellers as a whole much better off. We’ve really focused at a high level on doing two things really well. One, make it much easier for people to find great products on Etsy. And two, once they’ve found those products to actually buy them. 

With 62 million products for sale, picking for any given buyer the 20 or 30 that should be on page one of search results is a pretty interesting and pretty challenging task. There’s been a lot of advances in machine learning that take things that would have been literally impossible ten years ago and made those things much more possible today. The key is understanding what an item is with relatively little data and then being able to determine for each individual person how to personalize search results. We’ve made leaps and bounds in the science of search and machine learning. That’s more relevant at Etsy than almost anywhere else.

The mission of Etsy is incredible. As the nature of work changes creativity can’t be automated. The role we play for creators and makers being able to harness their creative passions and power and turn that into a way to earn a living for their families is a mission that I think is ever more important in this fast-changing economy.