AI Case Study
StitchFix designs clothes for underserved customer segments using genetic algorithms
StitchFix, a US based online styling service, identified customer segments which are underserved and uses genetic algorithms to develop new designs to serve them better. They use a combination of selection combination and mutation mechanisms to identify mix of designs to be sent to human designers for final approval.
Consumer Goods And Services
"Stitch Fix has combined the expertise of personal stylists with the insight and efficiency of artificial intelligence to analyse data on style trends, body measurements, customer feedback and preferences to arm the human stylists with a culled down version of possible recommendations. This helps the company provide its customers with personalized style recommendations that fit their lifestyle and budgets.
Using the data it collects, the company is designing its own styles known as Hybrid Designs. They think of each style as a collection of attributes such as color, arm length and neckline. Then, they look at the feedback that’s available for each of these attributes. By recombining attributes and even mutating them slightly, Stitch Fix is able to create new designs to share with its human designers to vet the final styles that make it into their inventory. Then, the styling algorithm will get the new products into the hands of customers and when they share their feedback, the cycle of evolution continues.
We approach this opportunity with inspiration from genetic algorithms: we use recombination and mutation along with a fitness measure—the same mechanism used by mother nature in evolution by natural selection.
The first step is to think of each style as a set of attributes ("genes").
Then consider our vast set of styles this way, and consider the client feedback ("fitness") we have available for each of them.
Now consider creating new styles by recombining attributes from existing styles and possibly mutating them slightly. "
R And D
Stitch Fix is an online subscription and personal shopping service in the United States founded in 2011.
StitchFix has 2.2 Million active customers and has a share of >3% in online apparel market
"Genetic algorithms: recombination and mutation along with a fitness measure.
The first step is to think of each style as a set of attributes ("genes"). Then consider StitchFix's vast set of styles this way, and consider the client feedback ("fitness") available for each of them.
Now consider creating new styles by recombining attributes from existing styles and possibly mutating them slightly. Note that the number of possible combinations is very large (∏ki)."
"Clothing attributes, Client feedback"