top of page

Amazon has developed a technology for virtual fitting clothes

Amazon, the world's largest online retailer, is developing virtual fitting rooms to help shoppers order the right size. The company is already inviting the most active users to its office in New York to begin conducting 3D scans of their bodies (test subjects will be more than compensated). So Amazon wants to create a computer model to understand how shapes of different sizes can change over time.

If the technology is successfully tested, online shopping will no longer have to worry about whether this or that thing will sit exactly on your figure. It will be enough to upload a couple of photos, and the computer brain will understand all your sizes, and even make predictions for the future. The company hopes that this will reduce the number of returns, and allow it to finally defeat Zara, Macy's and GAP in the clothing business.

Enthusiasts are required to fill out a large online questionnaire before undertaking a 3D scanning procedure. That way Amazon can figure out what their diet is, how often they go to the gym, and if they have any weight loss plans. Those selected for the program will then be scanned every two weeks for 30 minutes. In total, they will need 10 visits, for each of which they will be “given” an Amazon gift card for $250. It is expected that during this time the system will “train” and be able to understand what size clothing should be recommended to customers for maximum return. For example, young children need to project things “for growth”, and guys who have just started going to the gym may need to leave more free space in the sleeves and shoulders. So Amazon hopes that the number of returns of clothes after an online order can eventually be reduced to zero.

Virtual fitting rooms are handled by a new department at Amazon, formed after buying startup Body Labs in October 2017. This company has developed a technology that captures a person from several angles and creates a realistic 3D model of his body, where virtual clothes are then superimposed. In November 2014, a startup from New York received the first $2.2 million to implement its technology, and already in October last year, Amazon, imbued with the idea, bought it in full for $70-$100 million (the exact amount was not disclosed then).

Now the Body Labs department has grown, and its ambitions have also become more serious. They have already learned how to create “3D statistical models of human bodies,” which are then combined with videos or photos of people using deep learning algorithms. That is, the system is “fed” with images of subjects three to five months ago. And they force her not only to create a 3D model from several photographs, but also to perfectly predict what has now happened to the shape of the subject's body. Who has increased the priest, who pumped up his bitsuha, who grew up, who lost a little weight. Going forward, Amazon users will be able to upload their photos or videos (most likely via a mobile app), and the technology will filter out suitable clothing for them based on their personal information. Farewell, size tables and attempts to calculate something on your own with a measuring meter.

In addition to marketing, Body Labs' work can be applied to games and augmented reality applications. For example, customers of the Amazon online store can be offered electronic “mirrors”, and fashion designers can immediately see how their new design will look on a person.

The fight against online "returnees"

The return of online shopping is a huge problem for online stores. By the way, we at Pochtoy fully understand them. You spend money on packaging and shipping the product, and then you have to return the entire cost to the buyer who did not like something, and again place the product in the warehouse, check it, repackage it, deliver it a second time. Last year, retail lost $400 billion (!). Even worse, this trend is growing: in 2017, 53% more goods were returned than in 2015. People begin to understand that it is convenient, and they will not get anything for this, because stores need satisfied customers.

If offline stores have about 8% returns on the corner (according to CBRE), then e-commerce has up to 15% or 30%. Almost all sellers notice this menacing growth, and do not want to put up with gigantic losses. According to an online survey, 44% of stores admit that they have to raise the price of goods to compensate for the cost of "returns". And 70% expect that they will have to devote even more resources to this in the future.

ZozoSuit version 1.0

But not everyone is ready to put up with this state of affairs. A week ago, we talked about a new development by the Japanese - the ZozoSuit suit, with which they want to become leaders in online clothing sales. They don't quite go in the same direction as Amazon. Their idea is to give each (!) of their customer a special suit that takes all the necessary measurements. More than 2 million suits have already been delivered free of charge to Japanese users, and at least 8 million more are planned to be distributed. But Amazon believes that this technology will not take root in America: it is much easier to upload a photo of yourself than to turn into a diver at home and pose for cameras.

Although something tells me that measurements from a real suit will always be more accurate than even the smartest computer model...

far-reaching plans

Amazon already accounts for more than half of all online electronics purchases in the US. Here it is one larger than Apple, Walmart, Costco and a dozen other giant corporations. But when it comes to clothing, the lead isn't that big, even after buying Zappos for $1.2 billion in 2009. Last year, Bezos (or rather, his right-hand man, Jeff Wilke, who makes all the decisions about the online store while the boss develops his Blue Origin) wanted to correct this shortcoming. Now the company will be even more associated with buying clothes online. To do this, Amazon has launched more than seventy of its own fashion brands in recent years, and also rolled out the Prime Wardrobe project, which allows subscribers to choose up to 10 items of clothing and shoes, get them for free, try them on, and then send the package back, and pay only for those things which they left.

The company loses tens of millions of dollars on Prime Wardrobe every year, but it is worth it, in its understanding, to accustom customers to ordering clothes on Amazon. And if the project with Body Labs is successful, there will be no need to return goods, and all these customers will become satisfied and profitable users.

Last April, Amazon also released the Echo Look device, a smart assistant inside the camera that helps with trying on clothes. Alexa sits inside it, taking pictures of you or filming your video in a new outfit. The perfect combo with Prime Wardrobe to instantly choose what to keep. You don’t need to hold a smartphone in one hand to take a photo, and machine learning algorithms tell you which outfit looks best. You “feed” the device two photos in different clothes, and through the smartphone application, it shows how much it likes one of them more - based on current trends, color, style, expert opinions. Over time, based on user responses, the self-learning system gets smarter and smarter.

Amazon's strategy is starting to pay off. The company already sells more clothing than Macy's. Own brands were the top sellers in sportswear and exercise wear in 2018 (accounting for 57% of Amazon's top-selling Activewear products for men and 72% for women). The fastest-growing brand is Amazon Essentials, which sells super-cheap clothes of decent quality. In the first quarter of 2018, it rose from #16 to #6 in popularity on the site.

Morgan Stanley expects Amazon to become the largest apparel and footwear retailer in the United States this year. Now he is in second place, with 7.9% of all sales. On the first - Walmart with 8.7%. The introduction of a new convenient online technology for trying on clothes will definitely attract customers. With it, Amazon can potentially leave competitors behind, and even reduce its own costs.

But for now, new technology is still on the horizon. Professor Susan Ashdown of Cornell University, an expert in the field of visualization of clothing and 3D scanning of the human body for its production, says that:

This technology is much more complex than many imagine. Your posture, how you hold your shoulders, the angle of your hips, all of these affect how well your clothes fit. How can a computer system evaluate this - perhaps only Amazon is able to crack this code. I think that the system is ready for the market, this is not a task for one year. But if they can, it will certainly be the beginning of a new era in online shopping.

P.S. For clothes, by the way, in Russia there are the most cheats. Especially children's things and things of famous brands, here the situation is more serious than with gadgets. We recently had a funny case with a football uniform. In the US, the uniform of the Russian team costs $ 89 (5,700 rubles at the current exchange rate). And at the World Cup, held in Russia, the Russian football uniform - from 8990 rubles, $ 50 more expensive. Even if you deliver one T-shirt from the States, it turns out to be more profitable. Russian football uniform for the Russian championship. From USA. To be honest, it's kind of surreal.

Well, you can always order from Amazon and other stores from We give a free address for you in the US to ship the goods. And we deliver them to Russia at the lowest price - from $8.99 per pound. And if you register with us with the Geektimes code, $7 comes to your account.

Tags: amazon, pochtoy, clothes, internet, shop, AI, 3D

Hubs: company blog, E-commerce management, Artificial intelligence

5 views0 comments


bottom of page