Pinterest is on a crusade to be a personal stylist for its 250 million monthly users, and it just introduced a total overhaul of its product pin system to do it.
Now, any time a user is on any home or style pin — say, for example, an image of a black silk shift dress — they will see a “recommended” section of items that are visually similar, and for sale. The user can then select to see even more options on a shopping-only feed.
“Pinterest has always aspired to feel hand-picked for you, recommending what to make for dinner or recommendations for your home; that has always been the vision — and with recent advances in machine learning and computer vision, it is a closer reality than ever,” said Pinterest spokesperson Jamie Favazza.
Despite advances in computer vision, engineers still have to methodically catalogue data on items; a specific type of bag, for example, might have a thousand similar styles, she said. In 2014, Pinterest acquired image recognition and visual search startup VisualGraph, and since then, it has accelerated its visual search capabilities. Lens, for example, launched in February of last year to let users match Pinterest content to a picture taken in “the real world.” Users can take a picture directly within Pinterest or upload a photo later.
Now, Favazza said, Pinterest is starting to unify text search and image search, meaning someone could do a text search for “holiday party style” and then upload a picture of a black dress, and Pinterest will ideally serve up visual inspiration for how to style that dress for a holiday party.
Favazza said the goal is not for users to be aware of using “high tech,” but rather to create a seamless, intuitive experience. “We don’t want the user to think, ‘I am going to type a query and then take a picture.’ Pinners are using very advanced machine learning and computer vision, and don’t even realize it,” she said.
To that end, the visual search engine this week is replacing its Buyable Pin format in exchange for Product Pins, which are designed to work better for both retailers and users. The new pins will be shoppable with real-time pricing and stock information (meaning no more “out of stock” dead-ends), and they will link out directly to a retailer’s site. When an item becomes out of stock, it will now no longer have a shopping tag icon, and it will ultimately be filtered out in exchange for newer products. Today, there are “hundreds of millions” of shoppable pins.
The Pinterest app also added a new shortcut called a long press that shows related, shoppable products from a variety of brands any time someone is looking at a home decor or fashion pin.
Retailers can enable Product Pins by adding specific metadata to their e-commerce product pages (which is free). When someone saves something from their site, Pinterest will automatically turn it into a Product Pin with dynamic pricing and stock information. Retailers can also work directly with Pinterest to add an entire product catalog to Pinterest as Product Pins.
A home decor pin and related, shoppable products, as seen on the Pinterest app
In an announcement on Tuesday, Pinterest’s head of shopping product, Tim Weingarten, reported that in the past quarter since Pinterest began testing the new shopping features, clicks on products to retail sites have increased 40 percent.
Influencers and small businesses can take advantage of shopping on Pinterest with a product-tagging feature called “Shop the Look.” Until September, this feature was only accessible to larger companies. The feature helps “democratize product discovery,” Favazza said. And, of course, any size brand can still pay for promoted pins.
Accessibility to smaller businesses has been an increasingly significant focus at Pinterest under the tenure of new COO Francoise Brougher, who came to Pinterest from Square in February. Brougher, speaking this week at the Marie Claire Power Trip held at Pinterest’s San Francisco headquarters, said that the changes this week are just some of many ongoing updates going forward.
A few months ago, the Pinterest app replaced categories, including a designated shopping category, with more “personalized and integrated recommendations,” based on user feedback that they wanted to discover shoppable items all over, in the same way they would find other recommendations on Pinterest.
Favazza emphasized that the updates simply reflect the way people shop.
“Instead of a dedicated space for shopping, you are shopping naturally, and a lot of times you don’t realize you are shopping. In tech, we call that ‘discovery’ or ‘search,’ but consumers call that shopping. We have taken an approach that integrates the experience and shoppable products with what people are doing on Pinterest,” she said.
She added that the more people use Pinterest, the more it will feel like its content is just for them. “The premise is to help you discover things you didn’t know you wanted before you even asked.”