Fifty-five percent of people who buy clothes online have returned an item because it did not look the same on them as expected (Google, 2023). As an online seller, you can avoid the same fate with Google’s new virtual try-on (VTO) tool.
The AI shopping feature allows users to visualize how clothing looks on different body types and skin tones – no more worrying about sizing or spending money on return shipping.
Starting June 13, Google Shopping users in the U.S. can virtually try on clothing from popular brands like Anthropologie, LOFT, H&M and Everlane. Simply click the “Try On” badge to access the feature.
The virtual try on technology is currently available for women’s tops, while the Men’s version will be released later this year.
What’s In Store for Shoppers?
Virtual try on technology isn’t new, but Google’s version is one of the first to offer realistic views of clothes on different body types using 3D models. These two new features offer a more enjoyable (and reliable) virtual shopping experience:
1. View Clothing on Different Skin Tones and Body Types
One of the biggest dilemmas of online shopping is seeing something you love but not knowing how it looks on you. In fact, 42% of consumers don’t feel represented by images of models. (Google, 2023)
Google’s generative AI shopping tool features real models, allowing users to see clothes on different body types as well as visualize how the color and pattern will look on their skin tone.
How is this possible? Google handpicked models in sizes XXS-4XL representing different skin tones, ethnicities, body shapes and hair types.
It gets better – users will also see how clothes accurately stretch, cling, fold, drape, wrinkle and form shadows on different body types at different poses.
“When you try on clothes in a store, you can immediately tell if they’re right for you,” said Lilian Rincon, senior director of consumer shopping products at Google.
“You should feel just as confident shopping for clothes online,” Rincon said.
2. Refine Your Product Search
Is the top you like out of budget? Or perhaps, the jacket fits well, but you don’t like the pattern. Online shopping just got easier with Google’s guided refinements feature.
With the help of machine learning and visual matching algorithms, shoppers can narrow their choices based on color, style and pattern. You can view items from multiple stores all in one place without the hassle of visiting each store separately.
What We Know So Far About Virtual Try On Tools
Virtual try on technology uses Augmented Reality (AR) technology, enabling customers to “try on” clothing and accessories without physically visiting the store.
Through AR clothing, consumers can easily compare products to determine which is right for them. This decreases the likelihood of returning an item that does not meet their expectations.
Virtual try on has been around for years, with notable brands like Converse testing the waters as far back as 2012.
Google itself has tested virtual try on tech in the past, collaborating with cosmetic brands like L’Oréal, Estée Lauder, Black Opal, MAC Cosmetics and Charlotte Tilbury. Shoppers were able to virtually test and sample different makeup shades on a range of models with varied skin tones.
How Google’s Virtual Try-On AI (Images) Works
Google’s virtual try-on technology is based on the diffusion model, a machine-learning algorithm developed by the company. The model gradually removes noise from an initial image consisting entirely of noise, taking it closer to a target through a step-by-step process (more on this later).
To ensure the highest accuracy, Google used images of people wearing clothes in various poses. For example, an image of a person wearing a shirt sideways and another of them looking ahead.
Furthermore, images featuring random garments on different people were fed into the system to eliminate visual flaws like fabric wrinkles and creases.
Understanding Diffusion-Based AI Models
To understand this model fully, let us explain Diffusion. Diffusion is a process in which noise or extra pixels gradually distort an image until it becomes unrecognizable. Subsequently, the noise is removed from the image, and the original image is reconstructed into perfect quality. This process is used in text-to-image models like Imagen to generate realistic images solely based on the text entered by the user.
Google took this process of diffusion and combined it with AR clothing technology. Instead of using text as the input for diffusion, they used a pair of images: one of a garment and another of a person.
Each image was sent through its neural network (a U-net), and information from both images was shared in a process known as “cross-attention.” This allowed for the generation of photorealistic images of people wearing garments.
Perfecting Google’s Virtual Try-On Tool
Google trained its VTO AI model using the Shopping Graph, a comprehensive dataset of products, sellers, brands, reviews and inventory.
As mentioned, the company used images of people wearing clothing in different poses. The AI shopping model then learned to match the garment’s shape from one pose to another.
This process was repeated again and again, using millions of randomly chosen images. Eventually, the model was able to generate realistic images of garments on different people from all angles such that users would be able to see what virtual clothing looks like on their desired model.
Why Should Brands Add VTO to Their eCommerce Strategy?
Beyond the convenience of trying on an item without waiting or leaving the house, virtual try on clothing offers other advantages.
Increased Sales and Customer Engagement
Virtual try on clothing offers customers a more interactive shopping experience. By allowing customers to try on items virtually, retailers provide a fun and engaging way for shoppers to explore different looks. This leads to more satisfied customers and repeat purchases.
Reduced Costs
A study by Barclaycard showed that 30% of shoppers return goods, while 19% order multiple versions of the same product to try and decide which one to keep. Virtual clothing helps reduce costs for customers and retailers by decreasing the likelihood of returns and avoiding unnecessary shipping costs.
Additionally, fashion eCommerce stores can repurpose fitting rooms to display more merchandise.
Gain Competitive Edge
The AR market is projected to be worth $198 billion by 2025 (Statista, 2022). Companies incorporating this technology into their operations can gain a competitive edge over other businesses and become industry pioneers.
With virtual try on clothing, customers can access a more immersive shopping experience that fosters long-term loyalty. They also encourage users to share their purchases on social media, creating an additional avenue for brand promotion.
Enhanced Customer Experience
Virtual try on clothes provides customers with a more convenient and personalized shopping experience. You can easily view items and get tailored virtual clothing recommendations and styling advice.
Increased Sustainability
Shipping and packaging costs are drastically reduced with virtual try-on technology. The fashion industry has increasingly trended toward sustainability in recent years, and virtual try on clothes helps brands and retailers align with this movement.
It also encourages customers to be mindful of their purchases by giving them a better idea of what an item will look like before buying it.
What an eCommerce Store Can Do To Adapt to AI Shopping
To keep up with the changing world of AI-assisted shopping, an eCommerce store needs to adapt its strategies. Here’s how:
Establish Clear Data Privacy Guidelines
Some consumers have privacy concerns with virtual try on apps collecting and storing their images and intimate measurements. Retailers must choose trustworthy technology providers to assure customers that their data is safe and secure.
Use High-Quality Product Images
Invest in high-quality product images that clearly show a garment’s shape, fabric and details of a garment. This will help customers accurately visualize what they are ordering, which is essential for making an informed purchasing decision.
Promote Inclusivity
Customers are likelier to trust companies featuring models that look like them. Retailers should strive for inclusivity by broadening their size range and featuring models from different backgrounds.
Offer Customer Support
As with anything new and unfamiliar, customers may have questions or need help using virtual try on clothes. Providing customer support via phone, email or chat is essential to set customers up for success.
Utilize Reviews Customers Reviews
VTO is still in its infancy. Encourage users to leave reviews about their experience with virtual shopping. This builds trust among new customers still on the fence and offers insights on improving AR clothing and similar technology.
Invest in Good Web Design
eCommerce hosting websites should be able to support AI-powered virtual shopping technology. Ensure your site loads quickly, offers a seamless checkout experience and provides informative product descriptions.
Work With an Expert on eCommerce for Fashion
Adjusting to a virtual shopping experience can be daunting. An expert on eCommerce for fashion can help retailers improve the virtual shopping experience so customers feel confident using the tool.
What VTO Means for the Future of eCommerce Shopping
With more fashion eCommerce businesses closing their physical locations and relying on digital sales, VTO is poised to revolutionize how consumers shop online.
VTO will allow customers to check out items quickly with a simple smartphone scan. This could reduce customers’ shopping time, leading to increased customer satisfaction.
Additionally, businesses can use the data collected from VTO eCommerce shopping transactions to better understand their customers’ preferences, allowing them to tailor their product offerings and promotions accordingly.
All these features hold the promise of making online shopping more efficient for both consumers and businesses alike. Note that Google’s new AI shopping tool is still in the early stages – there’s still more to be had from this emerging technology, and we’re here to witness it.
Thrive, a full-service digital marketing agency, helps fashion eCommerce stores leverage VTO to maximize their online shopping experience. We help businesses increase their sales through search engine optimization (SEO), Shopify web design product photography, eCommerce hosting websites and more.
Contact Thrive today to see how Google’s virtual try-on tool can benefit your business.
Frequently Asked Questions
HOW DOES GOOGLE’S VIRTUAL TRY-ON (VTO) TECHNOLOGY ENSURE USER PRIVACY, ESPECIALLY REGARDING THE USE AND STORAGE OF PERSONAL DATA AND IMAGES?
For privacy and data security, AI-powered virtual try-on technologies typically employ encryption and anonymization techniques to protect user data, ensuring that personal images and information are securely stored and handled.
CAN USERS CUSTOMIZE THE VIRTUAL MODELS FURTHER TO MATCH THEIR BODY SHAPE AND SIZE MORE ACCURATELY?
Google’s VTO features have the ability to display clothes on a variety of real models representing different sizes, skin tones, body shapes, ethnicities and hair types, utilizing the Monk Skin Tone Scale as a guide. However, we can’t ascertain whether users can customize virtual models further, such as adjusting height or specific body proportions.
HOW ACCURATE ARE THE VIRTUAL REPRESENTATIONS IN TERMS OF COLOR AND TEXTURE?
The accuracy of virtual representations in terms of color and texture can vary based on the technology used and the device’s display capabilities, with developers continuously working to improve faithfulness to the material.
HOW DOES GOOGLE’S VTO TECHNOLOGY HANDLE ITEMS WITH COMPLEX TEXTURES OR PATTERNS?
Google’s generative AI model for VTO, using a technique called diffusion, is designed to produce high-quality, realistic images of clothing on models. This includes accurately reflecting how clothes would drape, fold, cling, stretch and form wrinkles and shadows, which implies a capability to handle complex textures or patterns. However, the specifics of handling these complex designs were not detailed, indicating a focus on achieving a realistic portrayal over specific handling techniques.
ARE THERE PLANS TO EXPAND THE VIRTUAL TRY-ON FEATURE TO INCLUDE ACCESSORIES OR FOOTWEAR IN THE FUTURE?
IThe initial rollout of Google’s VTO technology focuses on women’s tops from various brands and mentions an upcoming extension to men’s tops later in the year. While there was no specific mention of plans to include accessories, footwear or other types of clothing, the technology’s design allows it to scale to more brands and items over time.