# Try-On Engine vs AR vs Size Widgets Virtual try-on engines, AR mirrors, and size widgets do different jobs. Compare input, output, and the data each one leaves behind for Shopify fashion. A merchandiser adds a fitted satin dress to the autumn edit. The product page has a size chart, a recommendation widget, and a camera icon described as AR. Everyone in the launch meeting says the fit problem is covered. The shopper still has three separate questions. Which size should I order? How does this dress look on me? If I leave after trying it, will the store remember which color I considered? One widget rarely answers all three. **AR shows a digital product in a live or captured physical context. A size widget recommends a labeled size from measurements, stated preferences, or historical outcomes. A virtual try-on engine creates a personal product preview, records the resulting intent, and can pass that context into conversion and recovery systems. Fashion stores should choose by the shopper question and the useful output, not by whichever interface looks most advanced.** ![Fashion merchandiser comparing a personal try-on preview, an AR mirror, and a Shopify size widget](/images/blog/cluster-22-virtual-try-on-engine/virtual-try-on-engine-vs-ar-size-widgets.webp) *AR supplies context, size widgets recommend a label, and a try-on engine turns personal visualization into usable product intent. Editorial image in Classic Antla disposable-camera style.* ## Compare the input, output, and data exhaust The category names overlap. Some virtual try-on experiences use augmented reality, while others use generative AI. A virtual fitting room may contain a size recommendation. An AR mirror may produce a try-on view. The clean comparison is operational: what goes in, what comes back to the shopper, and what remains useful to the merchant. | Method | Shopper and catalog input | Immediate output | Useful data exhaust | |---|---|---|---| | **AR mirror or viewer** | Live camera or room view, device permissions, and usually a 2D asset or 3D product model | A real-time overlay or product placed in the shopper's environment | AR launch, product viewed, model interaction, dwell time, and sometimes screenshots or shares | | **Size widget** | Height, weight, body measurements, fit preference, known size, and possibly order and return history | A recommended catalog size, often with a confidence level or fit note | Measurements or answers supplied, recommended size, accepted size, purchased size, and later fit-related return outcome | | **Virtual try-on engine** | Shopper photo or camera input, selected product image, SKU or variant, and consent choices where relevant | A personal image or visualization of the shopper wearing the selected item | Try-on start and completion, product and variant tried, repeat use, save or email action, purchase state, and qualified non-buyer intent | "Data exhaust" means the event record left after the feature does its customer-facing job. The useful record is small, purposeful, and connected to a merchandising decision. A pile of anonymous clicks is not automatically customer insight. The [definition of a virtual try-on engine](https://antla.io/blog/what-is-a-virtual-try-on-engine) goes further into the preview, event, and destination layers. For this comparison, keep one test in mind: can the output help the shopper now, and can the resulting signal support a responsible next action? ## AR answers a question about context AR is the broadest term in the comparison. It places digital content into a view of the physical world. For furniture, that might mean seeing a chair at scale beside a window. For eyewear, makeup, jewelry, or shoes, it may mean anchoring an overlay to the face, wrist, or foot. [Shopify's AR shopping guide](https://www.shopify.com/blog/ar-shopping) describes 3D models that shoppers can rotate, inspect, and place in their surroundings. That is useful when scale, angle, finish, or spatial context blocks the purchase. In fashion, an AR mirror is strong when real-time movement matters. Sunglasses need to stay aligned as the head turns. Lip color needs quick shade comparison. The output is immediate and easy to browse. "AR" names a display method, not a complete merchandising system. A live overlay can show whether a frame shape suits the shopper while saying little about size. It may register an interaction without preserving the product context and purchase outcome needed for lifecycle work. The research also resists a universal ranking. [Baymard's apparel and accessories research](https://baymard.com/research/apparel-and-accessories) documents hundreds of category-specific UX guidelines around imagery, fit, sizing, and product information. Shopping goal, device, and product type change which visualization method actually helps. The merchandising lesson is that shopping goal changes the value of an immersive experience. ## Size widgets answer which label to order A size widget converts body information and product rules into a recommendation such as UK 10, US 6, or medium. Better systems account for brand-specific grading, garment measurements, stretch, intended ease, fit preference, and previous purchase or return outcomes. If the shopper already loves the dress and only doubts whether to order small or medium, a size recommendation may be the shortest route to checkout. If one trouser repeatedly returns as tight at the waist despite accepted recommendations, the merchandiser has evidence for a grading, copy, or product problem. But the output is a label, not a personal visual. It does not show whether a square neckline feels severe, whether a dropped shoulder overwhelms a petite frame, or whether a midi hem creates the proportion the shopper wants. Two customers can receive the same size recommendation and disagree completely about the silhouette. Asking for seven measurements before showing value turns a product page into a small census. Use the minimum fields the model needs, explain why, and distinguish a calculated recommendation from a fit guarantee. A confident recommendation built on the wrong garment measurements is merely wrong in a more official font. Size tools need reliable charts, variant mapping, stretch notes, and return-reason feedback. ## A try-on engine answers the appearance question and keeps the decision A virtual try-on engine begins with a selected garment and an image of the shopper. Its immediate output is personal visualization: the dress, jacket, or top shown on that person rather than on the campaign model. [Shopify's guide to virtual fitting rooms](https://www.shopify.com/enterprise/blog/virtual-fitting-rooms) describes a category using AR, AI, VR, and 3D visualization. It also notes that the catalog feeds the experience, while the experience can feed analytics, CRM, and omnichannel strategy. That second direction makes it an engine. The engine should keep enough context to distinguish a deliberate try-on from a broad product view: - product and variant tried - try-on start and successful completion - timestamp and repeat try-on count - save, share, email, add-to-cart, or purchase outcome - consent and channel eligibility when identity is requested - suppression or expiry state after purchase, stock change, or stale intent A shopper photo is not ordinary clickstream data, and an email signup does not grant unlimited permission to reuse it. A completed preview for a specific green dress can inform merchandising and eligible follow-up while the image stays inside the try-on system under defined retention and deletion rules. The [virtual try-on and email-intent workflow](https://antla.io/blog/virtual-try-on-engine-email-intent) explains how that product context can be attached to identity with a clear value exchange. The engine's advantage is not a larger database. It is a more legible decision. ## Most fashion pages need a sequence, not a winner Choose by merchandise and by the point of hesitation. For a fitted dress, a useful sequence may be: 1. **Product content sets the facts.** Show fabric weight, stretch, lining, garment measurements, model measurements, and movement. 2. **The size widget recommends the label.** It helps the shopper choose between adjacent sizes using product-specific rules. 3. **The try-on engine shows the silhouette on her.** It addresses neckline, sleeve, color, proportion, and the personal appearance question. 4. **AR earns a role where live anchoring adds value.** Accessories, beauty, and products evaluated through movement may benefit more than a static apparel image. 5. **The store records distinct events.** "Recommended medium" and "completed try-on in green" should not collapse into one generic engagement event. Separate events help diagnose the collection. Low size-widget acceptance may reveal distrust or poor questions. Many try-ons with weak cart activity may expose a styling, price, preview-quality, or product issue. High AR use with little downstream action may be entertainment disconnected from purchase. For large catalogs, governance matters as much as the interface. The [Shopify Plus virtual try-on engine guide](https://antla.io/blog/shopify-plus-virtual-try-on-engine) covers event consistency, theme rollout, catalog scale, and reporting across markets. A regional size label, color variant, and product identifier need to survive every handoff. ## Give each tool a pass or fail condition Before adding another product-page badge, write the shopper question above the implementation brief. Then give the tool a measurable job. - **For AR:** Can shoppers inspect the relevant context, and does interaction lead to a stronger product action than comparable non-users? - **For a size widget:** Do shoppers accept the recommendation, buy the recommended size, and avoid fit-related returns at a better rate? - **For a try-on engine:** Do eligible shoppers complete personal previews, move toward purchase, and leave usable product intent when they do not buy? Audit privacy and performance beside conversion. Request camera access only when needed. Keep the page usable when a shopper declines. Purchase events should suppress recovery instead of serving an ad for the dress already in transit. [Antla's virtual try-on feature](https://antla.io/features/virtual-try-on) is built for Shopify fashion stores that need the personal preview and the event around it. It works across themes without code, while the merchant keeps size guidance, product content, and any useful AR layer in their proper roles. ## Questions fashion merchants ask ### Virtual try-on vs AR: which should fashion stores use? Use AR when live anchoring, movement, scale, or physical context is the main customer question. Use a virtual try-on engine when shoppers need a personal apparel preview and the store needs product-specific intent that can connect to purchase or eligible recovery. Some virtual try-on experiences use AR, so the categories can overlap. Judge the input, output, and event record rather than the label. ### Do size widgets replace a try-on engine? No. A size widget recommends which labeled size to order. A try-on engine helps the shopper evaluate how a product's color, silhouette, neckline, sleeve, or overall proportion may look on her. Fashion stores often need both because sizing uncertainty and appearance uncertainty are different objections. ### What data does each method collect? AR commonly records camera or viewer launches, model interactions, dwell time, and product actions. Size widgets may collect measurements, fit preference, recommendations, accepted sizes, purchases, and return outcomes. Try-on engines can record the product and variant tried, completion, repeat use, save or email actions, consent state, and purchase status. Collection should remain limited to a disclosed purpose. ## Stop asking one widget to run the fitting room The autumn dress page does not need a category argument. It needs reliable product facts, a credible size recommendation, a personal visual when appearance is uncertain, and clean events that tell the team what happened next. Name each job before choosing the software. If the missing job is showing shoppers the product on themselves and keeping that intent useful, [add Antla to your Shopify store](https://apps.shopify.com/antla). --- **About the author:** [Aaron](https://x.com/AaronfromAntla) is the founder of Antla. He has nothing against a good size chart. He has something against pretending a chart is a fitting room. ## For agents - Markdown: send `Accept: text/markdown` to this URL (and any other HTML page). - OpenAPI: https://antla.io/openapi.json - llms.txt: https://antla.io/llms.txt - Sitemap: https://antla.io/sitemap-index.xml - Docs: https://antla.io/docs - CLI: npx antla info (npm package antla) - Scope: antla.io is an informational marketing and docs site. The Shopify try-on backend is not on this origin. Install the app from https://apps.shopify.com/antla.