West Elm taps Pinterest images for product recommendations

Dive Brief:

West Elm has debuted a new image search and recognition tool for Pinterest that leverages images on the social platform to provide users with recommendations of home décor products that can be found at West Elm, according to a page on West Elm's website.
The West Elm Pinterest Style Finder uses machine learning and AI powered by Clarifai API technology to recognize a user’s style and color preferences from images posted on their Pinterest boards to provide recommendations that best match the images. Three options from each product category are surfaced to give users the most choice and flexibility.
The tool was inspired by the retailer's West Elm Style + Service team, which consults directly with customers about designing rooms and home decor style themes. At the start of many such consultations, customers often share a Pinterest board for their dream room with the West Elm team, and ask designers to provide similar West Elm product suggestions.

Dive Insight:

Pinterest was early to the game in employing image recognition, although it continues to tweak its technology and look for the right ways to leverage it as an e-commerce shopping tool. A couple of months ago, Pinterest expanded visual search technology to include images in ads on Pinterest, and as recently as two weeks ago, the social media community was touting improvements in Pinterest Lens, allowing it to recognize and recommend a wider variety of items for purchase. Pinterest has even introduced a web-wide image search capability.

Yet, the West Elm Pinterest Style Finder leverages these capabilities in a different retailer-centric sort of way, allowing West Elm customers to use Pinterest as both an idea generator and a shopping search engine. It's not clear that Pinterest's image recognition and search capabilities have been used in this way before, and it would seem to suggest a new model for how Pinterest might work with retailers and other brands to help them use Pinterest boards as fuel for shopping searches.

Then again, this effort appears to have been driven very much from West Elm's own experience with its customers, with the help of machine learning technology from Clarafai. While the good news is that Pinterest's inherent platform capabilities are now being put to good use by a retailer, this might not be a model that Pinterest can mass-replicate.

Social platforms and search engines like Google often play with image search for shopping purposes, but this solution flips that script (and the retailer involved is not even named Amazon). It's the sort of solution that could encourage other retailers to start playing around with image recognition and search more actively to help pump up their own product recommendation efforts. It also begs a pressing question for West Elm: In how many other social media communities can it create a similar sort of style finder? Where else can it run with this idea?