The AI-powered software that boosts sales by delivering ideas to consumers with automated “collections.”
Some ideas take several iterations to come to fruition. That’s the story of Michelle Harrison Bacharach.
Her New York startup, FINDMINE, has proven to boost total sales by as much as 10% for online retailers like adidas, Cubavera, Perry Ellis, John Varvatos and more. Yet those kinds of wins took Bacharach plenty of time — mainly finding out what the product really looked like.
While getting her MBA from NYU, she initially set out to create a consumer app that would help people find clothes online that would go with outfits already in their closets. But then Bacharach discovered it would involve taking photos of every single item in a person’s closet— and even she herself wouldn’t do that as a shopper.
Then, she realized that on the flip side, this technology could go through the inventory and images of retailers and solve a big problem for them. Merchandisers and wardrobe designers may spend hours creating outfits that work together, identifying the jeans that works with the scarf and the cool floral jacket, for instance.
So Bacharach started FINDMINE in 2014 to create artificial intelligence that would go through a retailer’s inventory and automatically create collections for consumers. The tech could offer consumers better ideas about how to wear the scarf they just purchased in different seasons, and with a range of seasons. Looking to match the new couch with accessories and pillows? FINDMINE could do that, too.
The result has been dramatic. In A/B tests of collections put together by a human designer and those created through FINDMINE’s AI, shoppers (and merchandisers) couldn’t tell the difference between the two. “It was hundreds of times more efficient and was good as they were,” says Bacharach.
Plus, the AI could create millions more smart outfits than a human ever could on their own. Those tests also showed that consumers spent three times more when they interacted with an outfit through AI. What are the KPIs? When customers interact, FINDMINE led to a 40% boost in average order value, a 44% higher conversion rate, and 130% more units per transaction. Of course not every shopper interacts with an outfit, so the total revenue contribution to the business is a 4% to 9% increase in site-wide revenue. “For a $100 million company, that translates to an extra $4 million to $9 million in revenue,” she says.
“They see the jacket, they see the boots and here’s how you can wear it,” says Bacharach. “People come back again and buy more frequently. There is a higher conversion because, ‘Hey, now see myself in that outfit.’”
The technology platform also reduces the chances of what Bacharach calls “latent resentment” toward a brand because a brand item never made it out of the closet and was forgotten. Says Bacarach: “With FINDMINE, brand affinity goes up because now they have the confidence to wear that item.”
For retailers, the API is simple to incorporate. The technology relies on computer vision technology and natural language processing and algorithmically sifts through inventory to match outfits that fit the brand vision.
Adidas and others now use the FINDMINE API in any number of 15 channels, including email campaigns that follow up with buyers to offer ideas about what to pair an item with or, even months later, how to wear it for a different season. The app also can be integrated into in-store iPads, above a rack of sweaters, for instance. Luxury brands can arm their store associates with style guides that help them assist customers inside the stores. By understanding what customers bought in the past, and turning to the vast AI-assisted collections, store associates offer even more personalized service.
FINDMINE now delivers 2 billion looks per year to its clients. Says Bacharach: “It’s really powerful, because it answers that question, ‘What do I do with that product?’”