way music streaming services have leveraged usage data to change how consumers discover new songs and artists, the potential for fashion to translate garment data and shopping behaviors into powerful algorithms has the potential to inspire great change in the $2 trillion apparel and footwear industry.
On an episode of The Impact Podcast
called “What Happens When Machine Learning and Fashion Converge?
” Jon Prial interviews True Fit Founding President and CEO Bill Adler on the ways in which shopping for clothing is changing, how data and machine learning are the future of shopping, and how retailers can connect people to clothes they’ll love, instead of ones they’ll return.
Bigger Data. Better Recommendations.
Creating smart enough algorithms to guide consumers to the right styles to fit both their bodies and their unique preferences is no small feat. “What’s unique isn’t really the algorithm,” says Adler. “What’s unique is the data set... By the end of this year, we’ll have 70 percent of the global fashion market. That’s 70 percent of a $2 trillion market.”
“What’s unique isn’t really the algorithm, and it’s certainly not computing power… What’s unique is the data set. It’s what’s underneath. AI is powerful, and it’s driven by better data.” – Bill Adler, True Fit CEO
What does True Fit's massive data set mean for both retailers and customers? It’s the difference between recommendations that resonate – and ones that get returned. “If I want to understand what outfits to create for you, I have to understand two things really well,” says Adler. “I have to understand your preferences, personally, and I have to understand all of the different attributes of the things that we think you’re going to like.”
Finding New Favorites
Each item possesses multiple garment-based data points, including measurements, style details, colors and more, but when paired with unique user profiles is where figures translate to personalized recommendations. Shopper behaviors are a powerful tool in making knowledgeable recommendations across different brands and retailers. “On the consumer side, we have about 400 million unique profiles: 400 million people who have bought something that is on our platform,” says Adler. “From that, we’re able to use AI to drive the preferences of that person.”
“If you bought and kept and wear the shirt that you’re wearing, and you love it… there’s something about that go-to item. These are the ones we want to learn about and structure the preferences to the person who wears them successfully.”
– Bill Adler, True Fit CEO
Using consumer profiles, designer data points and a powerful algorithm, generating the kind of suggestions that work for shoppers is where it comes together. “That’s where the magic of great software comes from,” says Adler. “We can show very seamlessly, very simply, the things that you’re going to love and stop showing the things you’re not going to care about.”
A Win for Designers
Providing fit data is one way brands can keep their product relevant and promote better sell-through on retail sites – without compromising proprietary data. “We’re a trusted third party in between the designer, the retailer selling that design and then the consumer trying to buy it,” says Adler. “Designers share data in a way that’s simple, and we convert it to structured, normalized data that allows us to provide recommendations for them on retailer sites.”
“If Michael Kors shares a dress’ information with us, when you’re shopping Macy’s for the Michael Kors dress, you know much more about the dress. Therefore, sell-through at Macy’s is higher, returns are fewer, the designer has a better relationship with the retailer and the consumer… and it didn’t cost them a nickel.”
– Bill Adler, True Fit CEO
Additionally, for the over 10,000 brands selling in today’s marketplace, many sell exclusively via wholesale and have access to limited insights on their consumers' behaviors. Working on the True Fit platform gives brands know-how beyond simple sales numbers. “We have a dashboard of analytical insights that we provide back to designers, which is pretty interesting stuff,” says Adler. “We are able to tell them who bought it, the demographics of that person, and we’re able to tell them granular things like attributes by sales, as in, these are your sales by sleeve length.”