The formula and the common variants
| Metric | Formula | What it tells you |
|---|---|---|
| Store conversion rate | Purchases divided by sessions, times 100 | How well the whole store turns visits into orders |
| Add to cart rate | Sessions with an add to cart divided by sessions, times 100 | Whether product pages convince shoppers |
| Checkout conversion | Orders divided by checkouts started, times 100 | Whether shipping, payment or forms lose buyers |
| Product page conversion | Orders of one product divided by views of its page, times 100 | Which products, photos or prices underperform |
Shopify's guide to conversion metrics gives the same core formula, online store conversion rate equals purchases divided by total sessions times 100, and explains why sessions are used rather than unique visitors: one shopper can visit several times before buying. Count orders, not items: an order with three products is still one conversion.
What moves conversion in a clothing store
Fashion buyers cannot touch the fabric or try the size, so the product page has to answer those questions. The levers most stores test first:
- Product images. A clear main image, the garment on a person, details of the fabric and a consistent look across the catalog. See our ecommerce product image guide.
- Fit information. A size chart, the model's size and height, and honest notes such as runs small.
- Trust and delivery. Shipping cost and returns policy visible before checkout.
- Speed. Heavy images slow the page on mobile, where most fashion browsing happens.
Change one thing at a time and compare the same period, otherwise you cannot tell which change moved the number.
Do AI product photos improve conversion?
There is no reliable public number for this yet. On 2026-11-12 we looked for studies that measure how AI generated product photos change conversion, and found no peer reviewed study we could read in full. Vendor blogs quote lifts, but without their method or sample, so we do not repeat them here.
What the research does show is that photos carry the product page. In Baymard Institute's usability testing, 56% of users' first action on a product page was to start exploring the product images, and Baymard found that 25% of e-commerce sites provide images that are not good enough for shoppers to explore and judge the product. AI photos can only help conversion through the same route: more and clearer images of the product, on a person and from more angles, than a store could shoot before.
The largest field study of generative AI in online retail we found, by Fang, Yuan, Zhang, Donati and Sarvary, ran randomized experiments on a cross-border e-commerce platform and reports sales effects from no detectable impact to 16.3% depending on the workflow. It tested text and service tools such as a pre-sale chatbot and product descriptions, not product images, so it says nothing direct about AI photos.
The honest way to know is to test on your own store: replace or add AI images on a group of similar products, leave a comparable group unchanged, and compare conversion over the same weeks. Check every AI image against the real product first, because a photo that promises a different colour or fit can raise returns even when it raises sales.
Conversion in a virtual try-on funnel
If you add Wizia Virtual Try-On to product pages, its analytics show a separate funnel for shoppers who use it: clicks on the try-on button, completed try-ons, add to cart actions after a try-on on the same page, and purchases detected on the order confirmation page. The conversion figure shown there is the ratio of try-ons to purchases.
The numbers come from the shopper's browser through the widget, so ad blockers and unusual confirmation page addresses can undercount purchases. Watch trends over a week or two rather than single days, especially with low traffic. Images for the product pages themselves come from AI Studio.