Virtual Try-On and Size Recommendation: Two Different Problems
Virtual try-on helps answer “how might this look on me?” Size recommendation helps answer “which available size should I order?” A generated image is not a measurement or a physical fit test. A suggested size does not show whether someone likes the garment's appearance. A fashion store can offer both experiences in the same app.
ChatGPT’s new Try on button, launched on 1 October 2026, shows the same split: it generates a preview, and OpenAI says the images do not guarantee fit or size. See what ChatGPT virtual try-on means for Shopify fashion brands.
Rendered Fits offers AI virtual try-on, size recommendation and Complete the Look on Shopify product pages. Its current Shopify listing, checked 14 September 2026, describes sizing informed by measurements and order outcomes. This guide explains how to choose and evaluate the two functions; it does not claim that either has a universal conversion or returns advantage.
What Is Virtual Try-On?
Virtual try-on presents a visual preview of a shopper wearing a product. In a photo-based experience, the shopper provides a suitable photo and the system combines it with product imagery to generate a result.
That preview can support a style decision: colour, outfit combination and overall appearance. Assess whether it preserves the real product's print, neckline, closures and silhouette. A persuasive image that changes a distinctive design detail can create the wrong expectation.
Do not infer a precise hem position, fabric behaviour, physical comfort or size from a generated image alone. Those need product information and fit evidence. Technology varies: photo-based generation, live overlays and 3D simulation are different approaches, with different inputs and limits.
For brand presentation and visual review, see our virtual try-on guide for luxury fashion brands.
What Is Size Recommendation?
A size recommender interprets shopper information and available sizing evidence to suggest a product size. Inputs can include measurements, fit preferences, size charts or purchase and return outcomes, depending on the provider. Not every system uses every input, and not every recommendation includes a calibrated confidence score.
A useful result is tied to the garment and its available variants. A general “you are a medium” profile is less informative when a brand sells fitted shirts, relaxed knitwear and structured coats with different intended fits.
Ask what happens when measurements conflict, product data is missing or the suggested variant is unavailable. Those are evaluation questions, not a claim that every app already handles them correctly. Our Shopify size recommendation page explains how to prepare a catalogue for Rendered Fits evaluation.
Side-by-Side Comparison
| Factor | Virtual try-on | Size recommendation |
|---|---|---|
| Main question | How might this look on me? | Which available size is suggested for me? |
| Typical inputs | Shopper photo and product imagery for photo-based tools | Shopper inputs and product sizing evidence; methods vary |
| Output | Visual preview | Suggested size and, where supported, fit explanation |
| Critical data quality | Accurate, clear product images | Correct chart type, units, product mapping and size labels |
| What it cannot establish alone | Physical fit, comfort or exact dimensions | Whether the shopper likes the visual result |
| Useful quality check | Review garment details in generated images | Compare suggestions with known fit cases |
| Business evaluation | Usage, purchases and mature outcomes | Usage, recommended-size selection and mature outcomes |
Setup effort depends on your catalogue, integration and data quality. Compare current plan terms directly rather than assuming an entire technology category has one price or implementation time.
Which Problem Are You Actually Solving?
If customers are returning because of appearance/style mismatch
Review the actual customer explanation. “Not as expected” might mean colour, fabric feel, construction, length or size. A visual preview could help with some of those questions, while better photography or fabric descriptions could be the more direct fix for others.
Use try-on in a small pilot where appearance uncertainty is documented. Check that the output represents the product accurately before measuring shopper response.
If customers are returning primarily because of incorrect sizing
Review exchanges and messages alongside return codes. “Too small” might identify an inaccurate chart, an unsuitable cut or a preference for more room. Fix known chart errors before evaluating a size advisor. Keep the chart available so shoppers can understand and challenge the suggestion.
Rendered Fits includes sizing as well as try-on. Evaluate its size guidance on the products you know well, including between-size cases and different cuts, before a wider rollout.
If you don't know which problem is bigger
Sample recent support conversations, size exchanges and returns with enough time for the return window to close. Label size, appearance, product quality and unknown reasons separately. Keep overlapping causes visible; do not force each return into one explanation when the evidence supports more than one.
Prioritise the most clearly evidenced friction. There is no universal rule that fashion stores should buy try-on first or sizing first.
Body Measurements, Garment Measurements and Ease
A body chart describes the person a size is intended for. A finished-garment chart describes the item itself. The difference between a body dimension and a corresponding garment dimension can include wearing ease or design volume; stretch garments need additional interpretation.
For example, a 100 cm jacket chest and a 92 cm shopper chest differ by 8 cm. That arithmetic does not establish that the jacket fits: you still need the measurement method, shoulder and sleeve requirements, layering and intended cut. A 50 cm flat chest width is also not the same quantity as a 50 cm chest circumference.
This is an illustrative data check, not a Rendered Fits sizing formula. Use the size-chart checklist to label units, distinguish body and garment values and map tables to products.
When You Need Both
Both functions can be useful when shoppers need a size decision and a visual decision. A tailored jacket, for example, requires attention to dimensions and intended fit while also raising questions about colour and outfit styling. Keep those answers distinguishable in the shopper journey.
The Shopify listing describes Rendered Fits as combining photo-based try-on, AI sizing and outfit features, with interface customisation and analytics. Treat that as a feature description; a particular merchant's fit accuracy or financial result still needs its own evaluation.
For a premium catalogue, use the luxury fit evaluation worksheet to assess product fidelity and sizing separately. Avoid displaying two competing recommendations without explaining which product and fit preference each refers to.
The 5-Minute Decision Framework
- Name the shopper question. Is it appearance, size or both? Use actual support and return evidence.
- Inspect your inputs. Check product photography and product-specific size charts before adding automation.
- Choose representative products. Include different cuts, materials and size boundaries, not just easy examples.
- Check the whole journey. Review mobile interaction, suggested variants, basket selection and visible product information.
- Set an evidence threshold. Separate a generated image, a used recommendation, a purchase and a kept item in your reporting.
To evaluate business impact, compare appropriately matched or randomly assigned groups where practical, account for promotions and stock changes, and wait for returns to mature. Influenced revenue is not automatically incremental revenue. The sizing and try-on measurement guide provides the definitions to use.
Frequently Asked Questions
What is the difference between virtual try-on and size recommendation?
Virtual try-on provides a visual preview. Size recommendation suggests a product size from shopper information and sizing evidence. They answer different questions and can be offered together.
Can virtual try-on also tell me which size to order?
An app can include both features, as Rendered Fits does. The generated try-on image itself does not establish a size; the size recommendation is a separate output to evaluate against product data and fit feedback.
Which has better ROI — virtual try-on or size recommendation?
There is no universal winner. Compare the problem each solves on your store, its full cost and measured outcomes. Count returns after the return window and distinguish purchases associated with usage from additional purchases caused by the feature.
Do I need both virtual try-on and size recommendation?
Consider both when customers have appearance and sizing questions. Start with a limited product selection and evaluate each function separately. No return-rate threshold automatically proves that you need both.
Does size recommendation replace a size chart?
No. Keep accurate sizing information accessible. A recommendation helps a shopper interpret available evidence; it does not remove the need for accurate charts or guarantee comfort.
Does size recommendation require 3D garment models?
Requirements depend on the system. A measurement-based size advisor does not inherently require a 3D model. Confirm the provider's product-data requirements, including chart type and supported categories, before committing.
Try Rendered Fits on Your Products
Review the current Shopify listing, watch the product demonstration and check current plans and allowances. Bring a small set of products, reliable size information and known fit cases so you can judge the experience against your own catalogue.