Returns in fashion e-commerce often look like a warehouse, courier, and customer support problem. In reality, they start much earlier: in the customer’s mind, at the moment they click “Buy now.”
When we buy shoes, a bag, or a jacket online, we’re deciding based on flat product images, studio lighting, and imagination. Then the package arrives—and what’s real loses to the version we filled in ourselves. That’s why returns are largely a confidence problem, not just a logistics problem.
Virtual Try-On (VTO) doesn’t “fight returns” the way return policies do. It does something smarter: it reduces the number of purchases that were set up to disappoint from the start.
Across retail, returns are a major financial drag. National Retail Federation (NRF) reports that in 2024, total returns in the US were projected at US$890 billion, and retailers estimated that returns would account for 16.9% of annual sales.
In fashion, this hurts even more because returns aren’t an “exception”—they’re baked into the shopping model. ICSC reported that the average online return rate was about 3× higher than in brick-and-mortar (15.2% vs 5%). And for apparel retailers specifically, consumers returned 22% of online purchases (vs 6.2% in stores).
That means even if reverse logistics runs flawlessly, the business is still paying for something that—in a perfect world—wouldn’t happen at all.
The key observation is simple: many returns don’t happen because customers “changed their mind,” but because reality didn’t match expectation.
In a PowerReviews study on apparel return behavior, the most common reasons were very “human”:
These aren’t “I changed my mind” reasons. They’re signals that the product page simply didn’t give the customer a clear enough picture of what they would actually receive.
ICSC also notes that online returns often involve: damage, poor fit (“did not fit”), and situations where an item was “not as expected.” More importantly, in their study, 87% of people who “over-order” online do so in apparel—specifically to try items at home and send some back.
In short: if shoppers can’t evaluate a product “on themselves” or in real scale, they start shopping as if their home is the fitting room.
VTO flips the default dynamic of online shopping. Instead of asking the customer to mentally fill in missing information, it shows the product in context—something traditional e-commerce struggles to provide.
Research on AR product experiences helps explain why this works: AR can reduce uncertainty about product quality and fit by increasing “informativeness,” a sense of presence, and how easily shoppers can imagine the product. (sciencedirect.com)
Footwear: proportion on the foot, the shoe’s silhouette, styling—and “is this even my vibe?”
Bags and backpacks: size perception (is it mini, medium, or “half your life in there”), how it sits on the body, and proportion relative to the shopper’s silhouette.
Apparel: increasingly, VTO supports the style decision (how it looks “on me”), but there’s an important caveat: historically, many solutions looked more like a visualization than a trustworthy fit simulation, which limited confidence.
The most important “side effect” is a good one: once shoppers can see the product on themselves, in true-to-scale context, the need to buy “just to try” drops.
In the field (outside of sales decks), a consistent pattern shows up: engaging with AR/VTO before purchase is often associated with lower returns and higher conversion—though the size of the effect varies by industry, category, and the quality of execution. (shopify.com)
Shopify describes an AR implementation example where adding AR coincided with, among other outcomes, a 5% reduction in return rate (alongside conversion lift).
In accessories, it’s especially clear that “scale + context” can be a big deal. In a case study by Tangiblee for The Little Green Bag (selling bags and accessories across multiple countries), one product reportedly had a 40% return rate before deployment, dropping to 20% after implementing visualization/try-on tools (a halving for that item). (tangiblee.com)
At the level of a major fashion player, Zalando reports that in pilots starting in 2023 with selected brands—including Levi's—they observed return-rate reductions of up to 40% in the jeans category, which they position as a basis for scaling in 2026.
It’s also worth being honest about where some of the big numbers in the industry come from. Trade sources often compile broader “proof points” (e.g., lower returns among AR users vs non-users), but the underlying data is frequently reported by vendors and partners—not always as fully independent, peer-reviewed meta-analyses.
The practical takeaway is still consistent: VTO can reduce returns when it genuinely changes the purchase decision—and that hinges on experience quality and whether shoppers actually use it.
Classic fashion e-commerce has a built-in “trap”: when we’re unsure about size, cut, or scale, many of us behave rationally…and expensively for the retailer:
“I’ll order two sizes (or two options) and return one.”
Economists and shopping-behavior researchers have a name for this: bracketing—buying extra sizes/variants to manage fit uncertainty, with an explicit plan to return some items later. This mechanism is discussed, for example, in research on bracketing in the context of free returns and size uncertainty. (papers.ssrn.com)
This lines up neatly with consumer data: both PowerReviews (13% of returns linked to buying multiple sizes) and ICSC (87% of over-ordering happening in apparel) show that “test-buying” is real and common.
VTO targets the root cause: it reduces uncertainty (fit / look / scale) before payment. AR product-experience research explicitly links reduced fit uncertainty to more favorable product attitudes in online shopping.
This matters, because the industry has already lived through multiple waves of “wow, we have virtual try-on!”—and just as many disappointments. Industry analysis suggests many earlier implementations stalled at the pilot stage due to issues like accuracy, latency, and the cost of scaling to thousands of SKUs, plus a simple reality: users don’t always want to turn on a camera and “perform” a try-on in every shopping context. (theinterline.com)
From a UX perspective, that leads to a straightforward rule: if VTO is going to reduce returns, it has to be all of the following at once:
Research on consumer acceptance of VTO technology (an extended TAM model) highlights the role of perceived usefulness, ease of use, decision comfort, and trust in adoption. (emerald.com)
In short: if VTO is going to reduce returns, it can’t just be “pretty”—it must be trusted, used, and scalable.
At WEARFITS, we look at returns as a purchase-decision problem: the customer should “understand the product” before they buy it. That’s why we focus on the elements that directly drive confidence, adoption, and scale.
The first pillar is realism and credibility. AR research suggests the key is reducing uncertainty by increasing “informativeness” and presence—which, in practice, means visual quality that doesn’t feel like a filter. (sciencedirect.com)
The second pillar is scale and operational practicality. In fashion, it’s not enough to build a “hero experience” for a handful of products—because returns happen at volume. That’s why it matters to deploy broadly without manually modeling every SKU. In one of our product articles, we describe an approach where products can be transformed from studio photography into 3D models (e.g., from a set of images), reducing dependency on expensive design files and making scale easier.
The third pillar is the “decision moment” and UX friction. In practice, VTO works when it’s easy enough that users don’t feel like they’re doing an “extra step”—they just naturally sanity-check the product. This aligns with VTO adoption research: ease of use and usefulness shape attitudes and intention to use.
Treating returns purely as a reverse-logistics problem is like putting out a fire without checking where the sparks are coming from. In fashion, those sparks are often uncertainty—which turns into over-ordering or disappointment after delivery.
There are environmental consequences too, which make returns more than “just a financial” issue. Research on fashion e-commerce returns highlights the sheer scale of costs and emissions (including estimates of UK returns costs and 2022 CO₂ emissions). (sciencedirect.com) Meanwhile, an EU analysis of returned-apparel flows suggests a significant share of returned items may never reach another consumer, and that emissions embedded in producing and distributing unused returns can dominate the footprint of return transport itself. (sciencedirect.com)
That’s why VTO is so compelling commercially: if it helps customers “get it right” on the first try, it reduces returns without restricting the right to return—simply by improving the decision upfront. And that’s the most scalable way to have fewer returns: not by making returns harder, but by building confidence.