After two years of manually curating ‘Shop the Look’ offerings, Pinterest has launched its new and improved fully automated experience.
The Pinterest Engineering team has announced changes to its ‘Shop the Look’ feature in its latest blog post, revealing its new fully automated experience for iOS users.
“Every day, people come to Pinterest to discover new ideas — and when they find a product they want to buy, it should be easy to purchase it. In home decor specifically, millions of people come to Pinterest to find inspiration,” the company wrote in its post.
Originally launched in 2017, the company had been using its human-in-the-loop approach to match products with pins. This was problematic because it prevented the image sharing platform from scaling to meet the needs of its millions of users.
“[We] needed a better way to scale across the billions of images we show Pinners. As a solution, we used computer vision to fully automate the process of matching products to scenes.”
According to the business, the ‘Shop the Look’ update will increase the coverage of each applicable Pin by up to 22.5 times across billions of Pins and products. Early testing has reportedly lifted engagement by seven percent.
Ninety-seven percent of the 1,000 most popular searches on Pinterest are currently non-branded. However, as part of Pinterest’s push to improve its shopping capabilities, the business says it will look to tag organic pins that have not been linked to a business account. Meanwhile, products that have been listed for sale from a specific retailer will go through the social platform’s revised three-step process to guide shoppers from pin to purchase.
Pinterest built its new technology by focusing on data collection, machine learning modelling, and serving. By collecting data, Pinterest is actively training its machine learning models and preparing its systems to identify and localise home décor objects in a Pin. The automated process will classify objects into product categories and then digitally represent them with similar images. These images are then linked with products and query Pins (the look) during the serving stage where visual embeddings are used to find the closest product matches to a specific Pin.
“This update brings more computer vision-powered results across Pinterest, showing visually similar ideas to more people. With more Shop the Look Pins in the system, Pinners can expect to see a much more consistent user experience across all home decor scenes,” Pinterest says.
“In the long term, the scene images are great resources to learn the relationship between objects, i.e. what objects complement each other or go well together in a certain style. We hope to leverage this rich data of object occurrence and build a sophisticated object graph for every object in the world, making Pinterest a personalised stylist for home, fashion and more.”