Inside Snap’s New Division to Help Brands With AR and AI-powered Shopping
Years of development and acquisitions by Snap Inc. have all led to this: On Thursday, the Snapchat developer unveiled its new augmented reality for enterprise services, or ARES (pronounced like “Aries”) for short, to take its AR technology beyond its Snapchat to brands’ and retailers’ own sites and apps.
The tech that makes this possible is Camera Kit, a developer tool set that the company has been working on for years. In recent months, Snap expanded the AR kit to shopping and now, it’s wrapping support and services around it to form ARES’ debut offering, Shopping Suite.
More from WWD
2020 Holiday Gift Ideas: Google's Top Searched Beauty, Home and Tech Gifts
How "The Bachelor" Contestants Have Gone From Reality T.V. to Mega-Influencers
The key features of the suite take direct aim at some of the top pain points for brands in immersive forms of shopping: creating 3D assets for the AR features, whether virtual try-on or through a 3D viewer; offering fit and sizing recommendations, and getting support to make it all work, and not just with each other, but also with whatever back-end system is already in use.
“[It’s] something that we’ve been preparing for for many years,” Carolina Arguelles, global AR product strategy and product marketing lead, told WWD. “It is the ability for us to take the world-class AR and AI technologies that we’ve been pioneering on Snapchat — and reaching such a massive audience with — and translating that, combining the AR technology with really strategic acquisitions that we’ve made over the last few years.”
Indeed, one can draw a direct line to the offerings from Snap’s acquisitions of Fit Analytics, the artificial intelligence-powered fit and sizing platform; Vertebrae, which specializes in creating 3D visuals for retail, and Forma, developer of photorealistic avatars. For Arguelles, assembling those capabilities amounts to something of a dream team to formulate Snapchat’s AR and AI-powered shopping.
The new software-as-a-service division, under the leadership of Jill Popelka, Snap’s head of ARES, is focusing on retail first with the Shopping Suite, aimed at fashion apparel and accessories, as well as home furnishings.
“Over the last decade, we’ve been hard at work bringing fun and personal AR experiences to Snapchatters,” she said. “In the next decade, we’re excited to take our world-class AR technology to business’ websites and apps. We look forward to making the shopping experience more delightful for consumers and transforming businesses around the world with AR Enterprise Services.”
Arguelles broke it down further: Accessories, like sunglasses from Goodr, offer a live view so shoppers can try them on in real time. Others, like the sweaters on Gobi Cashmere, work by seeing the clothes overlaid on top of still images of preset models in different sizes and body shapes or on top of uploaded selfies — at least for now. There are distinct challenges in representing fabric and its physics across different body types, but Snap is working on it.
Even for static try-ons, it apparently takes a lot of tech and data science to make that work. “What our approach does, by passing it through 20 deep learning models, is it allows us to estimate how those pants would look if you’re slightly angled more if your hair is in front of your shirt. Where does that hair start in and what’s shirt and what’s hair. That’s really what this approach allows us to do,” Arguelles explained. “Then we’re able to then apply that to images of models that they want to see or of course the user’s uploaded photo.”
Retailers can offer a 3D viewer, so customers can see what the products look like and rotate them, to view different angles.
Notably, brands don’t have to invest in scanning equipment or other ways to capture their merchandise. Snap can take the same 2D product images they already use in catalogues, apps and sites and feed them into their AI-powered imaging system to pump out 3D versions.
Data science also plays into fit and sizing recommendations, which is the piece of the puzzle that the Fit Analytics team brings to the table. The so-called Fit Finder solution is exactly what it sounds like — a tool that helps customers find the right fit.
Tying it all together is the “Enterprise Manager” back-end. Snap describes that as “an infrastructure and back-end system to manage and create AR, making it simple to utilize our SDKs [software development kits], create AR experiences and manage 3D asset catalogs — all while seeing real-time performance analytics on how AR is driving consumer engagement and conversions.”
Think of it as a control panel or dashboard that wrangles all of those components.
It’s a system “tailor-made for retail that actually integrates the technology with these individual products that are sold, these product catalog integrations,” continued Arguelles.
So Camera Kit is an important basis for ARES and Shopping Suite, but there’s a lot more, she added. “What we learned is that a retailer needs a specialized customer support team that knows how to integrate with commerce back-ends, a retailer needs specific technology solutions for them and their problems that they’re trying to solve, such as fit and sizing in addition to AR try-on.”
There’s another dimension with ARES: Although brands can stick with the Snapchat app alone for some of these features, blending the tools into their own online, or even physical, businesses supercharges the capabilities, because data can flow in two directions.
Take the fitting features, for instance.
Arguelles explained: “In order for a retailer to integrate Fit Finder, we do need those data flows back and forth. On one side, we need to ingest their information about the purchases that are happening of that item and the returns, [such as] which sizes are being returned. We triangulate that with obfuscated data about the shoppers, so we know who purchased that item in a medium and returned it for the large. And we’re able to aggregate and use our AI algorithms to then make personalized recommendations for every single new shopper.”
Snap has a deep well of AR data. Of its 375 million daily active users, more than 250 million use AR in the app daily.
In other words, the information enables key features and makes them smarter. Meanwhile, Snap can feed the data back into the brands’ systems, which can inform other algorithms and efforts, like loyalty programs, future NFT initiatives and other offerings.
This is what Snap has been moving toward for a long time, and it seems to know that it can be overwhelming for some retailers. To help with that, the company developed hands-on integration services as well. Think of it as a white-glove service offer by an in-house team to help clients implement these solutions.
Brands like Goodr, Princess Polly and Gobi Cashmere have been testing these features for months now. Snap shared that Gobi Cashmere’s conversion rate was four times higher for shoppers who used its fit and sizing recommendations, and one in four shoppers try out the virtual try-on tool.
Best of WWD
