How Eileen Fisher is using AI to help customers find their best-fitting sizes
As more companies look to AI to help solve business pain points, Eileen Fisher is among the latest women’s apparel brands to apply it to fit in the hopes of driving down returns and increasing conversions.
Eileen Fisher, a 40-year-old New York-based brand, launched an artificial intelligence-powered sizing tool called The Closet in late March as part of its spring campaign. The tool uses eight real models with different body types that customers can select and use to “try on” items to help envision the fit and drape of a piece of clothing. For example, they can click to see what a shirt might look like from the back or tucked in. The tool uses AI to size and overlay the garment to each model’s body shape and size.
So far, The Closet has driven 50% more conversions than shoppers who don’t use the tool, Blair Silverman, vp of e-commerce at Eileen Fisher, told Modern Retail. It has also lowered the bounce rate and average order volumes are higher. Silverman credits this to people being more confident that their purchases will be a fit.
“It goes back to our mission of mindful shopping,” she said. “You’re really being mindful about your purchasing and being able to try it on.”
As artificial intelligence becomes a mainstream technological solution, brands are finding customer-facing purposes. EMarketer found that as of February 2024, about one-third of e-commerce marketers said they were using AI for product recommendations. But this is increasingly showing up toward fit and sizing. Both Amazon and Google have recently launched their own AI-powered virtual try-on or fits insights tool.
Elsewhere, some brands are beginning to toy with virtual sizing tools to personalize the product based on the user’s own measurements. Kickstarter brand Doubl makes made-to-measure bras using an app that takes measurements of someone’s body via the camera. An AI tool then calculates what the product measurements need to be. The brand sold nearly 250 bras during its Kickstarter launch in April.
Roger Williams, head of loyalty for relationship marketing firm Marigold, said it’s not hard to envision a future where major brands allow people to create AI-generated images of themselves to try on clothes virtually. That’s similar to what’s already being done at some retail companies, which are using augmented reality to show shoppers what items look like in their homes based on photos they upload of their rooms.
Williams said fit technology could potentially be one of the most profitable use cases of artificial intelligence in the e-commerce world because of the high cost and logistical challenges of returns.
“It’s being driven by the numbers because it’s such a big cost center,” he said. “As it grows, people get used to it, and at some point, they’re going to find it indispensable.”
But there will be hiccups. Models that use AI to make product suggestions require context, like if someone is buying a product for themselves or as a gift for someone else. “One of the opportunities retailers have with a fit finder as they’re going to get more and more popular is to be able to have some zero-party data context,” Williams said.
In the case of The Closet, the tool isn’t doing any predicting of what a customer should buy. Rather, the shopper can choose their own size based on what body type they are most aligned with.
Silverman and others who are proponents of fit technology say the key unlock is helping customers understand what a product will look like on them. While the “standard” model is 5’10” and wears a size extra small, data from The Closet shows people are more likely to look at models that are shorter, a size medium or fuller figures.
“That’s representative of the people that are shopping,” she said. “That in itself is a win for us because we’re able to give our customers a model that reflects them so they can shop with confidence and see how it’s going to look on their body shape and type.”
This story has been updated to reflect that Eileen Fisher is headquartered in New York. It has also been updated to explain that the tool uses real models.