September 25, 2019
Pinterest is now promoting itself as a “visual discovery engine,” where its 300 million global monthly active users can not only browse billions of images but also purchase the items they find there. For most of its users, Pinterest is a go-to place for home décor, gardening and personal style, allowing users to “pin” or post photos to create inspiration boards. According to Pinterest, its AI technology can accurately pinpoint 2.5+ billion objects in photos, millions of which can be purchased by clicking on the item.
The Wall Street Journal reports that Lens, an AI-powered feature, allows Pinterest users to zoom in on items in a photo. Lens “can automatically identify the object, provide a link to retailers that sell it and show photos of similar items that can be purchased” — plus, it “also works with photos that users upload themselves.” Then the user can click on any of these so-called shoppable items to go to the brand’s website to purchase it.
In 2014, Pinterest also acquired VisualGraph, whose technology helped boost its object-recognition capabilities.
Pinterest head of engineering Jeremy King noted that “when you’re building something that’s based on machine learning or computer-vision technology, what you need is a giant corpus of data.” “Pinterest is the master of that,” he said. “People come to Pinterest with shopping intent, and oftentimes in the past they had to leave Pinterest to go find the item. Now, we’re making it easier to get to the retailer.”
At eMarketer, principal analyst Andrew Lipsman said, “the technology could be critical to Pinterest’s monetization strategy … [because] it has been difficult for advertisers on Pinterest to truly understand the link between an impression, or a user’s view of a pin, and their purchase behavior.” Since Pinterest can now identify multiple objects in a single photo, “it can show more recommendations for users to purchase,” something easier for advertisers to understand.
In the quarter ending June 30, Pinterest had revenues of $261.2 million, up 62 percent from a year earlier, with global average revenue per user at 88 cents, up 29 percent. Due to the costs for its IPO, Pinterest also posted a net loss of $1.16 billion. For computing and storage, Pinterest relies heavily on Amazon Web Services, but its AI-enabled visual search engine is built in-house.