On this page
Google wound down Doppl, its standalone virtual try-on app, in 2026 — less than a year after launching it — while folding a "Try it on" experience directly into Search. For fashion brands evaluating try-on technology, the lesson isn't that try-on failed. It's that try-on only works when it lives where shoppers already are.
Doppl launched with real ambition: point a phone at an outfit, or a photo, and see yourself wearing it, powered by generative AI. It got press coverage, App Store placement, and a genuine technical achievement under the hood. Then, within a matter of months, Google discontinued the standalone app and moved the underlying capability into Search's shopping experience instead. That sequence — build a destination app, watch adoption stall, fold the feature into an existing surface — is worth studying closely, because it is not the first time this has happened in try-on, and it will not be the last.
Why Standalone Try-On Apps Struggle to Retain Users
A standalone app asks for something most shoppers won't give: a second download for a single use case. Even a well-built try-on experience competes against every other app on a phone for attention, and try-on is not a daily habit like messaging or social media — it's an occasional task tied to a specific purchase decision. That mismatch between app-store friction and infrequent usage is a structural problem, not a product-quality one.
Doppl's app also had to solve discovery cold. Users had to hear about it, install it, then remember to open it before buying rather than after. Compare that to try-on embedded in Search or on a product page: the shopper is already there, already deciding, and the try-on moment requires zero extra steps. Context switching kills conversion intent — a shopper who has to leave a retailer's site to open a separate app is a shopper who may not come back.
The Pattern: Try-On Wins When It's Embedded, Not When It's a Destination
Google's own move tells the story. Rather than shutting down try-on as a capability, it moved the feature into Search, where billions of product queries already happen. That's the opposite of retreat — it's a bet that try-on is more valuable as a layer inside an existing shopping flow than as its own app. Retailers should read this as validation of embedded try-on, not as evidence that try-on doesn't work.
This mirrors what's already true for size recommendation. Fashion brands that ask shoppers to think about fit as a separate task — a size chart in a PDF, a quiz on a different page — see far lower engagement than brands that put a size recommendation or virtual fitting room directly on the product detail page, next to the size selector, at the exact moment of purchase decision. The channel matters as much as the technology.
| Approach | Where it lives | Friction to first use |
|---|---|---|
| Standalone try-on app | Separate download | High — install, learn, remember |
| Try-on folded into Search | Existing search flow | Low — already in the moment |
| Try-on embedded on PDP | Retailer's own site | Lowest — zero extra steps |
What This Means for Fashion Ecommerce Brands Right Now
Return rates remain the underlying reason any of this matters. Apparel and footwear return rates sit well above other retail categories, and US returns totaled roughly $890 billion in 2024, with online return rates hovering near 19-20 percent according to the National Retail Federation. Try-on and size recommendation exist to attack that number, and the delivery mechanism determines whether shoppers actually use the tool often enough to move it.
Brands watching Doppl's shutdown shouldn't conclude that generative try-on is a dead end — Google didn't kill the capability, it relocated it to a higher-traffic surface. The practical takeaway is to evaluate any try-on or fit vendor on where the experience actually sits: embedded in your storefront, on the product page, inside checkout — not as a separate destination shoppers have to seek out. For deeper context on how the underlying models work, see our explainer on how AI try-on works with diffusion models.
Where YourSizer Fits
YourSizer is built as an embedded widget, not a standalone destination. It sits directly on the product page inside your Shopify or custom storefront, next to the size selector, so a shopper never leaves the retailer's site or opens a separate app to check fit. That design choice is deliberate: the Doppl shutdown is a live case study in why embedded context beats a standalone app, and it's the same principle behind how YourSizer's Shopify integration works — install once, and fit guidance appears exactly where the buying decision happens. Brands get both size recommendation and try-on visuals in one flow, without asking shoppers to install anything or think about fit as a separate errand.
Frequently Asked Questions
Why did Google shut down Doppl?
Google has not published a detailed public breakdown of Doppl's usage numbers, but the pattern is consistent with what happens to most single-purpose try-on apps: low repeat engagement relative to the cost of building and maintaining a standalone product. Rather than abandon the underlying try-on technology, Google moved it into Search, where the feature can reach shoppers already in a purchase-research mindset instead of requiring a dedicated app install.
Does this mean virtual try-on technology doesn't work?
No — it means the delivery format matters as much as the underlying model quality. Try-on technology, including generative and diffusion-based approaches, continues to improve and continues to reduce uncertainty at the point of purchase. What struggled was the app-as-destination model, not try-on itself. Embedded try-on on retailer product pages has a fundamentally different usage pattern than a general-purpose app, because it appears exactly when a shopper is already deciding whether to buy.
Should fashion brands still invest in try-on and fit technology?
Yes, and the Doppl episode actually strengthens the case for doing so correctly. Returns remain a major cost center — McKinsey has estimated returns cost fashion brands more than $40 billion globally each year — and any tool that reduces size-driven returns pays for itself quickly. The key decision is choosing a solution that lives inside your own storefront rather than asking customers to adopt a new app, which is the exact distinction Google's own strategy shift just validated.
How is embedded try-on different from what Doppl offered?
Doppl asked users to open a separate app, often outside the context of a specific retailer or purchase. Embedded try-on, by contrast, is installed once by the retailer and then appears automatically on relevant product pages, so the shopper experiences it as part of browsing rather than as an extra task. This difference in placement is often the single biggest driver of adoption and usage frequency, more so than differences in image quality between vendors.
Conclusion
Doppl's shutdown is less an ending for virtual try-on and more a correction in where it belongs. Google's decision to fold the capability into Search rather than kill it outright confirms that shoppers want fit and try-on answers inside the flows they already use, not as a reason to download something new. For fashion brands, the actionable signal is straightforward: invest in try-on and size recommendation that live on your own product pages, and treat any vendor's roadmap toward deeper storefront integration as a feature, not an afterthought.