A shopper does not decide to buy a jacket at the moment they press "try it on". They decided to consider it in a feed, decided to click on a listing, decided to stay on the page, and are now deciding whether it will look like that on them. Virtual try-on answers one of those questions very well and the others not at all, and a store that installs a try-on button expecting it to do the whole journey's work is asking one tool to be four.
This page is a map of the journey with try-on placed on it. It publishes no lift figure and no return-rate figure; which method of try-on suits which product is in AI vs AR virtual try-on, and the install is in how to add virtual try-on to your online store. This page is about where the two kinds of try-on sit and what each one is for.
Where does virtual try-on belong in the buying journey?
In two places, doing two jobs. Before the shopper arrives, on-model images made from the garment — the same product on the range's models, in the poses and crops the channel needs — do the trying-on for everyone: in the ad, the listing thumbnail and the product gallery, they answer "what does this look like worn?". On the product page, a try-on the shopper runs on their own photo answers the last question, "what does it look like on me?", which no gallery can. Discovery and consideration are served by the images; decision is served by the shopper's try-on; after the purchase, both are what the delivered garment is compared against.
What is virtual try-on, for a store?
Virtual try-on is any image in which a garment is shown on a body it was not photographed on. For a store it comes in two forms. Seller-side try-on dresses the garment on a chosen wearer — a catalogue model, or a model the store photographs and keeps — to produce the on-model images a listing needs, made once and shown to every visitor. Shopper-side try-on dresses the garment on the shopper's own photo, on the product page, made once per shopper and shown only to them. The first is production; the second is a feature. They use the same underlying operation and sit at different points of the journey.
The journey, step by step
Discovery. The shopper is not shopping; they are scrolling. What stops them is a worn image: the garment on a person, in a scene, at the crop the placement allows. Shopper try-on cannot exist here — there is no shopper yet — so this step is entirely served by seller-side images. The question answered: "is this the kind of thing I want?"
The listing. In search results and category grids the garment competes with its neighbours at thumbnail size. A consistent on-model image — the same wearer, crop and light across the range — reads as one catalogue rather than a jumble. The question: "which one do I open?"
Consideration. On the product page the shopper looks at the gallery: front, back, detail, a lifestyle frame, ideally the garment on more than one body. This is where seller-side try-on does its most detailed work, because the shopper is now comparing the images to what they imagine receiving. The question: "is this garment what I think it is?"
Decision. The shopper likes it and hesitates. The hesitation is almost always one of two things: "will it suit me?" and "which size?". The first is the question shopper-side try-on exists for — the garment on their photo, their proportions, their colouring. The second it helps with only indirectly, by showing the intended fit on their body; a size guide still does the measuring. The question: "on me?"
Purchase and delivery. Nothing happens to the images here, but everything that was shown is now a promise. The delivered garment is compared, consciously or not, to the image that sold it.
Returns and the second visit. A return is the journey running backwards: the garment did not match the image, or did not suit the body. Images that showed the garment truly reduce the first; a try-on the shopper ran reduces the second. A shopper who kept the garment is a shopper whose next visit starts at consideration rather than discovery.
The two kinds of try-on, by what they answer
The mistake is to treat them as substitutes. They are not: one is made before the visit for everyone, the other during the visit for one person.
- Seller-side images answer "worn how?". They can show the garment on several bodies, in every pose the range uses, at every crop the channels need, checked against the garment before publication. They cannot show the shopper's own body.
- Shopper-side try-on answers "on me?". It can show the garment on this shopper. It cannot be checked in advance, depends on the photo the shopper provides, and does only what the product page's own image lets it do — a poor product image produces a poor try-on for every shopper who tries.
- Both depend on the source. A garment photographed flat, sharp, complete and in its worn arrangement dresses well on a catalogue model and on a shopper. A garment photographed badly dresses badly on both.
In Photta the seller-side images come from the apparel studio: a flat lay or mannequin shot, a wearer from the catalogue or one the store uploads and keeps, and the on-model set. The shopper-side try-on is Photta Business: a widget on the store's own product pages where a shopper uploads a photo and sees the piece on themselves, for apparel and jewellery.
Try-on versus traditional product photography: two answers, not one replacement
The comparison comes up because both cost money and both end in an image of a garment on a body, but they answer different questions at different moments, and a store that treats one as a replacement for the other loses the question the missing one answered.
Product photography — a photographed shoot or an on-model image made from the garment — answers the shopper who has not yet chosen: what does this look like, how does it hang, what is the fabric. It is made once per garment, by the store, before anyone arrives, and it is the same for every visitor. Shopper try-on answers the visitor who has chosen a garment and is deciding whether it is for them: how would it look on me. It is made per visitor, on the product page, from the store's garment image and the shopper's own photograph, and it is different for everyone.
So the honest structure is both: the catalogue and the gallery are photography, kept accurate to the garment, because they carry the discovery and the comparison; the product page adds try-on where the decision is made, because it carries the last doubt. Replacing the photography with try-on leaves nothing to discover; replacing try-on with more photography leaves the last doubt unanswered. What they share is the garment's pixels: a try-on made from an inaccurate garment image shows the shopper an inaccurate garment on their own body, which is the most convincing wrong answer a store can give.
Discover, consider, decide
The loop this page's checklist serves, named so the other try-on pages can point at it.
- Discover — worn images made from the garment, in the ad's and the listing's formats, the same wearer and light across the range.
- Consider — the product-page gallery: front, back, detail, lifestyle, more than one body where the range has them, each image checked against the garment.
- Decide — the shopper's own try-on on the page, fed by a product image good enough to dress from, and a size guide beside it doing the measuring.
Try-on versus AI model photography: who is wearing the garment
AI model photography and virtual try-on are often named as one thing because both produce a garment on a generated body. The difference is who the body belongs to, and it decides which one a store needs where.
In AI model photography the store chooses the wearer — built once, kept for the range, dressed with every garment — and the images are catalogue images: made before the visit, the same for everyone, reviewed against the garment before they go live. In shopper try-on the wearer is the visitor, the image is made during the visit from their own photograph, and nobody reviews it before they see it; the garment image the store supplied is the only part of it the store controlled.
That asymmetry sets the priorities. For AI model photography the work is in the wearer and the recipe, so the range reads as one catalogue, and in the review, so no garment is misrepresented. For try-on the work is entirely upstream, in the garment assets the store uploads: a garment photographed fully open, plain, sharp and complete, because every shopper's image is generated from it and there is no second chance to check. A store can run either without the other; most fashion stores end up with the catalogue made by the first and the product page carrying the second, both fed from the same garment photographs.
Try-on placement checklist
| Step | The shopper is asking | What answers it | Watch |
|---|---|---|---|
| Ad or social post | Do I want this kind of thing? | A worn image in the placement's crop | Click-through on worn images against flat ones, per placement |
| Listing grid | Which one do I open? | A consistent on-model thumbnail across the range | Open rate per product, with the odd ones out in the grid |
| Product gallery | Is this garment what I think it is? | Front, back, detail, lifestyle; the same garment in every frame | Time on page, gallery views, zoom use |
| Hesitation | Will it suit me? | The shopper's own try-on | Try-on starts, completions, and add-to-cart after a try-on |
| Sizing | Which size? | A size guide; the try-on shows intended fit, not measurements | Size-related returns, separately from fit-related ones |
| After delivery | Is this what I was shown? | Images that were checked against the garment before publishing | Return reasons, by product |
| Second visit | Where did I leave off? | The same worn images, recognisable | Repeat purchase among shoppers who used the try-on |
Every "watch" column is a comparison inside your own store, before and after, on the same products. None of them has a value this page can give you.
One coat, one journey
A store sells a wool coat. Before the season it photographs the coat flat and on a mannequin, front and back, sharp, with the collar as it is worn, and dresses it in the apparel studio on the range's two models: a front three-quarter for the ad, a straight front for the grid, back and detail for the gallery, one lifestyle frame for the page. Each image is checked against the flat lay before it goes out.
A shopper sees the ad, opens the listing because the grid is consistent enough to compare, and reads the gallery. They hesitate on colour — the camel against their own hair and skin — and run the try-on on their photo. The coat is the coat they saw in the gallery, on them. They check the size guide, choose a size, and buy. The coat that arrives is the one in the pictures, because the pictures were checked; it is a colour they had seen on themselves, because they had. Two questions, two tools, one garment, no figure claimed for any of it.
Frequently asked questions
Should I add virtual try-on to my product pages?
If you sell apparel or jewellery, a shopper-side try-on answers the one question the gallery cannot, "on me?", and it does so only as well as your product image allows. Add it after the product images are in order, not instead of them; a try-on on a poor product photograph gives every shopper a poor result.
Is virtual try-on the same as on-model product images?
They are the same operation at different points of the journey. On-model images are made by the store from the garment, once, for everyone, and can be checked before publishing. Shopper try-on is made on the product page from the shopper's photo, once per shopper, and cannot be checked in advance. A store needs the first; a store selling fit- or colour-sensitive products benefits from the second.
Where in the funnel does try-on have the most effect?
At the hesitation just before purchase, when the shopper likes the garment and is unsure it suits them. Earlier in the journey the work is done by worn images the store made; later, by whether those images were true. This page publishes no figure for the effect; measure it on your own products before and after.
Does virtual try-on help with sizing?
Indirectly. It shows the garment's intended fit on the shopper's own body, which settles "oversized or fitted?" and "how long on me?", but it does not measure. Keep the size guide beside the try-on and track size-related returns separately from fit-related ones so you can see which the try-on moved.
What does a shopper-side try-on need from the store?
A product image good enough to dress from: the garment complete, sharp, on a plain background, in its worn arrangement. The shopper supplies the other half, a photo of themselves. The store controls only its half, which is why the product image is the first thing to fix.
What this page does not claim
That try-on raises conversion or lowers returns by any amount, or that either kind of try-on is right for every store. What it claims is that try-on is two things at two points of the journey, that the first is production and the second a feature, and that both are only as good as the garment photograph they start from.
To make the on-model images the journey starts with, see virtual clothing try-on; to put a shopper-side try-on on your own product pages, see Photta Business; to check what the images show before they go out, see the virtual try-on quality-control checklist.