Virtual try-on adoption and conversion impact: what 2026 pilots report

Virtual try-on technology has moved from experimental pilot to measurable business case in the past eighteen months. Retailers integrating AR fitting rooms into product pages now report conversion lifts ranging from 15 per cent to 94 per cent, depending on category, implementation and measurement method. Yet the gap between published vendor claims and independently verified results remains wide, and adoption outside large platforms is still limited.

McKinsey and Business of Fashion identify AI and digital capabilities, including virtual try-on, as the single biggest opportunity for fashion in 2026, while 55 per cent of luxury consumers report having used virtual try-on tools. The commercial pressure is clear: fashion return rates frequently exceed 25 per cent and can approach 40 to 50 per cent in fast-fashion segments, and US retail returns reached 849.9 billion dollars in 2025.

The technology addresses purchase hesitation more than post-purchase regret. Fashion e-commerce converts at roughly 2.81 per cent on average, below food and beverage at 5.29 per cent and beauty at 4.71 per cent. Virtual try-on does not fix sizing, but it does give shoppers a clearer preview of style, drape and colour before they commit.

What conversion lift do retailers actually see?

Shopify reported that merchants using 3D or AR product experiences saw a 94 per cent increase in conversion rates compared to standard product photography, a figure drawn from the platform's own merchant data and cited in multiple industry reports. That headline number sits at the top of a wide range.

Rebecca Minkoff found that shoppers who interacted with a 3D model were 44 per cent more likely to add an item to their cart and 27 per cent more likely to place an order, while customers who viewed a product in AR were 65 per cent more likely to make a purchase. The fashion brand reported these results after implementing AR for handbags on Shopify.

Retailers integrating virtual try-on for eyewear, accessories or outerwear reported a 15 per cent average conversion lift compared to standard product pages, based on A/B tests. Snap Inc. found that AR try-on experiences resulted in a 2.4 times increase in purchase intent, though the company did not disclose which brands or test conditions produced that figure.

The variation reflects differences in baseline conversion, category, device mix and how the lift is measured. A 94 per cent increase on a 2 per cent baseline produces a 3.88 per cent final rate; a 15 per cent lift on a 3 per cent baseline yields 3.45 per cent. Both can be true and both matter, but they are not directly comparable.

How much do return rates fall?

Return-rate reductions show a similar pattern: large ranges and limited disclosure of test conditions. Snap Inc. reported a 36 per cent reduction in return rates for fashion products using AR try-on. Zalando reported returns reduced by approximately 40 per cent in early virtual try-on tests, a retailer-reported figure that has not been independently audited.

Amazon Fashion launched AI-powered virtual try-on for shoes in 2024, expanding to clothing in 2025, with early results showing a 25 per cent reduction in shoe returns and a 35 per cent increase in customer satisfaction scores. Amazon ended its Try Before You Buy programme on 31 January 2025, citing the combination of limited scale and growing use of AI-powered features including virtual try-on and personalised size recommendations.

McKinsey research documented return-rate decreases of 25 to 40 per cent across 14 apparel brands using digital fitting technology, though the study did not name the brands or specify which technology vendors were used. Most published return-rate claims come from vendors or from pilot-stage retailer tests, not from sustained, scaled deployments.

Which product categories show the strongest results?

Eyewear, accessories and beauty consistently deliver stronger results than apparel. AR try-ons for glasses saw a 30 per cent increase in add-to-cart rates, and virtual makeup try-ons drove a 25 per cent increase in conversion for lipstick and foundation. Face-based AR works well because the tracking is accurate and the visualisation is close to what a customer would see in a mirror.

Rebecca Minkoff reported that 65 per cent of customers who used AR try-on for handbags were more likely to purchase. Apparel and clothing lead virtual try-on adoption by application segment with a 46.8 per cent share, driven by size doubt and costly garment returns.

Full-body apparel remains the hardest category. Rendering fabric drape, stretch and fit on diverse body types in real time requires either high-quality 3D assets or advanced generative models. Google introduced a custom image generation model for fashion in 2025 that understands the human body and nuances of clothing, including how different materials fold, stretch and drape on different bodies, working at the scale of billions of items in its Shopping Graph. Google shut down its Doppl app on 30 April 2026 and moved try-on into Search.

What did the June 2026 luxury shopper study find?

In June 2026, vendor DRESSX published an industry report examining the effect of virtual try-on on purchase intent and conversion among more than 1 million luxury shoppers across over 200 countries, finding that the technology increases purchase intent and conversion while reducing product returns. The company reported aggregate data from its own platform rather than independent retailer results.

Virtual try-on users convert 50 per cent more often than shoppers who never use the feature, according to the study. 44 per cent of virtual try-on users remained active 30 days after their initial visit, while the non-user retention rate was 1 per cent. Mobile devices drove 70 per cent of try-on interactions and accounted for more than four out of five revenue dollars.

These figures are vendor-published platform statistics and reflect high self-selection: engaged shoppers are engaged shoppers. The retention gap is likely overstated because the non-user baseline includes one-time browsers. Still, the directional signal is consistent with what other platforms report.

How big is the virtual try-on market, and who offers it?

The global market for virtual try-on was valued at 12.5 billion dollars in 2024 and is projected to reach 48.8 billion dollars by 2030, growing at a compound annual growth rate of 25.5 per cent. 71 per cent of consumers say they would shop more often if they could use AR to try products, yet only about 1 per cent of e-commerce stores currently offer it, a figure that counts all e-commerce businesses including micro-stores.

In January 2026, Zara introduced interactive virtual try-on experiences that allow shoppers to preview outfits digitally before purchasing, helping the brand reduce return rates while increasing buyer confidence. Snapchat partnered with Gucci in February 2026 to launch the company's first Sponsored AI Lens for a luxury fashion brand's ad campaign, powered by Snapchat's proprietary generative AI technology.

As of 2026, more than 250 million Snapchatters have engaged with AR shopping lenses over 5 billion times, and 80 per cent report feeling more confident in their purchase after trying a product on with AR; among daily Snapchat users specifically, 63 per cent say having a virtual try-on feature is key to driving their decision to buy.

What are the limits and open questions?

Virtual try-on does not solve fit. It shows how a garment looks, not whether it will fit a customer's measurements. Fit and size are the main reasons for online apparel returns, accounting for 42 per cent of returns, according to the National Retail Federation. AI size recommendation engines address that problem separately, and some vendors now bundle both.

The technology is expensive to deploy at scale. Full-body apparel try-on requires either 3D assets for every SKU or a generative model trained on garment behaviour. Small and mid-market retailers lack the budget or technical capability to build it in-house, and white-label solutions are not yet plug-and-play. Up to 90 per cent of AI initiatives fail to scale, McKinsey noted in its 2026 report.

Accuracy varies by lighting, device and body tracking. AR overlays work well on faces and hands but struggle with dynamic fabric and diverse body shapes. Generative models can produce realistic images but may not preserve garment details that matter to the brand. Fashion requires a high level of accuracy, precision and attention to details when it comes to AI in merchandising and try-on; brands care about drape, styling, silhouette, textures, fit and brand identity, not just a generic visual output, argued DRESSX co-founder Natalia Modenova.

There is also a measurement problem. Platforms publish aggregate lifts but rarely disclose test conditions, control groups or baseline metrics. Retailer pilots test different implementations in different categories on different customer bases. A 30 per cent lift for eyewear is not the same as a 30 per cent lift for dresses, and a vendor-reported figure is not the same as an audited result.

What about sustainability and cost?

Virtual try-on reduces the need for physical samples and minimises product returns, helping brands lower their environmental footprint. Each avoided return saves reverse logistics, restocking and markdowns. Returns cost US retailers more than 400 billion dollars a year, and AR try-on attacks the visual uncertainty behind many of them.

But the cost of deploying the technology is front-loaded. Creating 3D assets costs between 50 and 200 euros per SKU, and turnaround can take weeks per collection. Generative models are faster but require integration, testing and ongoing refinement. Smaller brands without developer resources struggle to justify the investment unless they see a clear path to payback within two to three seasons.

What are the risks if companies wait?

The technology is becoming table stakes for categories where visual preview directly impacts purchase confidence. McKinsey and Business of Fashion identify AI, including virtual try-on, as the single biggest opportunity for fashion in 2026, with 55 per cent of luxury consumers having used virtual try-on and 15 per cent using it frequently. Large platforms including Google, Amazon and Snapchat have embedded try-on into their shopping experiences, setting a new baseline for what consumers expect.

Retailers who delay risk losing share in categories where competitors offer a better pre-purchase experience. But rushing into deployment without clear success metrics or category fit also carries risk. The lesson from 2025 and early 2026 pilots is that virtual try-on works when it solves a real purchase barrier in a category where baseline conversion is low and return rates are high.

What should fashion companies do now?

1. Audit your return data by reason code. Identify which categories have the highest returns driven by style mismatch, colour expectation or visual uncertainty rather than fit. Those are your priority candidates.
2. Start with accessories or eyewear, not apparel. Face-tracking AR for glasses, watches and jewellery delivers stronger results with lower implementation cost than full-body garment rendering. Prove the ROI on a narrow use case before expanding.
3. Test on mobile, not desktop. 70 per cent of virtual try-on interactions happen on mobile devices. If your mobile product page experience is weak, fix that before adding try-on.
4. Require vendors to disclose baseline and control group data. Ask for test conditions, sample size, category mix and how lift was calculated. A vendor case study without those details is a marketing claim, not a business case.
5. Bundle try-on with size recommendation tools. Virtual try-on addresses style confidence; AI size engines address fit. Combining both produces better outcomes than either alone, especially for apparel.
6. Plan for content production, not just technology integration. 3D asset creation or generative model training takes time and budget. Factor that into your total cost of ownership and timeline.
7. Measure engagement, not just conversion. Track interaction rate, time on page, add-to-cart by user segment and return rate by try-on usage. The goal is not just higher conversion but better-informed purchases that stick.