Finding the right size bra online can be an exercise in frustration. That’s why intimates company Wacoal used to tell customers to give up.
“It is hard for women to size themselves or measure themselves,” Miryha Fantegrossi, vice president of merchandising and design at Wacoal America, told WWD. “It’s been a challenge for a very long time, and that’s why Wacoal has always encouraged people to go into a store and have a proper bra fitting.” Such a session acted as a starting point for a dialogue or exchange between the customer and the fitter, with the ultimate goal of finding the appropriate size, with the right coverage and support, and in looks that appeal to the individual’s taste.
Those days are gone now. On Tuesday, the company introduced MyBraFit, a mobile app for iOS and Android that uses artificial intelligence which, in combination with a short quiz, identifies the right size and style for the user. The app offers a selection of bras from which the customer can shop.
The app doesn’t require users to measure themselves — which tends to be error-prone anyway, according to Fantegrossi — nor does it even ask the user to enter the bra size they usually use. That’s by design, since most women don’t wear the correct size.
One reason is the lack of standardization in the bra business, which is a problem that has persisted for decades, and the mere fact that many women just don’t know how a bra is supposed to fit or where the straps are supposed to lie.
Wacoal’s app aims to take the guessing game out of the equation.
It’s far from the first to try replacing a human bra fitter with technology, though the attempts have varied. In some cases, the brand asks shoppers to upload selfies, while others focus on asking pointed questions about breast shape or other details.
MyBraFit combines those scenarios, blending computer vision with user-supplied information. The shopper takes a selfie wearing tight pants and a bra, and then the app asks four questions about tissue type, shoulder type, breast shape and the type of coverage she wants.
The company teamed up with a tech partner — Sizer Technologies, which specializes in mobile-based scanning — to develop the system. It scans and generates the user’s body measurements, while Wacoal uses the other information to inform its recommendation engine, which uses deep learning to match the person with choices from its range of bras. The selection spans 67 options and one goal of using the technology is to reduce the chance of recommending sizes that the company doesn’t carry.
One area that might put people off, though, is the selfie process itself. In the privacy era, the mere thought of uploading one’s photo in a state of undress could have a chilling effect on would-be users.
That spurs Wacoal to emphasize the steps it’s taking to protect user privacy. The selfies are “de-identified” — which means that the scans aren’t actually realistic photographs, but images that look more like a negative or X-ray, with the subject’s facial features blotted or obscured. So anyone worried about, say, a hacking event turning their near-naked selfies into the latest meme might rest easy.
The scan is transmitted with an anonymous identification number, allowing the system to track the data and analysis. It extracts the body measurements to provide sizing advice, then makes the data available to Wacoal.
According to Fantegrossi, no realistic photograph is saved anywhere — whether on the user’s own photo library, as a file transmitted to Sizer Technologies or in the data provided from the vendor to Wacoal. And the scan that’s used to pull the measurements gets destroyed, per the company’s terms. Users can also access their data and delete it, if they wish.
Such privacy policies are important across technology, but especially when it comes to the sensitive matter of selfies with women posing in their bras.
Some women won’t want to do it, regardless. But others might, and the more who do will help boost the app’s accuracy.
“It’s early…but I would say our initial results have shown at least a 300 percent improvement in accuracy over the sort of industrywide size charts that we and a lot of our competitors have used for years. So it’s a very significant improvement,” the executive said.
Part of the accuracy stems from the company’s work, measuring as many as 500 women over the past year, which was no simple feat during the pandemic.
“In order to capitalize on the level of information that you could get from a digital image capture, we had to revamp our size charts, so that they could speak to each other,” she explained. “They just existed, and they haven’t been touched by anybody in forever. This is where the advancement comes — because the customer doesn’t have to take her measurements, you don’t have to worry about how she would take it or what exact position she’s taking it from. And therefore, the information that we get is better.”
The other reason for the jump in sophistication is the advanced evolution of the smartphone camera itself.
While Wacoal doesn’t make use of technology like Apple’s LIDAR — the laser-based sensor in the rear camera of the latest iPhones, which offers a major upgrade in accuracy for digital measurements — the current front-facing selfie camera is still much better than those used in years past.
The demise of ThirdLove’s bra-fitting app features speaks to that evolution, on top of the other challenges Fantegrossi mentioned.
The app, which was built in 2013, demanded good photos to work well. But it was hard to get people to follow the detailed instructions for posing in the right lighting and framed in just the right way. Ultimately ThirdLove abandoned the app, going for an online quiz instead.
Of course, there are still others and now, with better cameras available, along with better AI and computer vision, apps like MyBraFit hope to ride the tech to the top of their mobile app category.