Pinterest is pushing its visual search engine for the real world - and it could upend the advertising business
And the company seems to have a clear message for marketers attending the event, something like: "If you think of us as an app where people post pictures of places they want to visit or things they want to buy for their living-room makeovers, think again. We're going to be the company that helps you use your phone to search the real world using images."
Indeed, Pinterest is putting a big emphasis on Lens and the visual search technology behind the product. The still-in-beta tool allows people to snap photos of physical items that bring up visual search results related to the items. Take a picture of a shirt in a store, and get back a series of images of similar shirts.
Pinterest placed Lens demonstration kiosks on the beach at Cannes, where people could snap photos of their own clothes or other products.
The company sees the tech as a major differentiator, one that may even help Pinterest expand beyond its core commerce-and-crafting audience.
The company's president, Tim Kendall, told Business Insider at the company's Cannes pavilion opposite the Carlton Hotel: "We think it's how discovery is going to increasingly be driven on the phone. We think it's going to be image-driven.
"We think it's a really big deal. I think our computer vision technology is the best in the world."
The Lens tech is also baked into Samsung's mobile browser, meaning many who don't use Pinterest will be able to learn more about images they encounter when out surfing the mobile web.
Pinterest recently raised $150 million and is now valued at $12 billion. So it's a good bet that investors will want Lens to pay off in revenue.
Ultimately, ads may come to Lens. But more immediately, Kendall sees the company's investment in visual discovery technology as significantly advancing search advertising, an area that Google dominates but Pinterest has recently made a big push into.
Kendall talked about how a theoretical furniture retailer might have to run ads tied to thousands of generic terms like "classic living room" to identify its products in the hope that they help people find the retailer in a web search. But Pinterest's machine-learning tech can automatically pull thousands of signals from images that can be used to deliver more relevant ads with far less labor, he said.
For example, that retailer's products may surface when people look for furniture items that fit with their specific home's design, Kendall said.
"I think this is the way that ads are going to be served," he said. "We are going to invest more in this. I think keywords are going to go away."