Clothing sells on its photos more than almost any other category, because shoppers cannot touch the fabric or try the fit. In Salsify's 2026 consumer research, 61% of shoppers said product images and videos are the biggest product-page factor when deciding whether to complete a purchase (Salsify, 20 January 2026). Baymard Institute's usability testing found that 56% of users' first action on a product page is to explore the images (Baymard, 2 April 2020).
So the question is not whether to invest in clothing photos, but which ones to take. Below are ten image types that each answer a specific question a shopper has, how to shoot each one, and where AI can produce it from a photo you already have.
The ten ideas at a glance
| # | Image | The shopper's question it answers |
|---|---|---|
| 1 | Front view | What does it look like? |
| 2 | Back view | What is on the back — fastening, print, cut? |
| 3 | Side view | How long is it, and how deep? |
| 4 | Fabric close-up | What is the material like? |
| 5 | Construction close-up | How is it made — seams, lining, buttons? |
| 6 | On-model view | How does it fit on a body? |
| 7 | Ghost mannequin | What is its shape and cut, without a model? |
| 8 | Flat lay | What is the whole garment, laid out? |
| 9 | Colour variants | What does each colour really look like? |
| 10 | In context | How big is it, and how would I wear it? |
1. Front view
The front view is the catalog image: the one used in listings, search results and category pages. Show the whole garment, uncropped, centred and consistently framed across the range, so a category page reads as one collection. Many channels set rules for it; Zalando, for example, accepts plain, lifestyle or ambient settings, applies its own light grey background to primary packshots submitted on white, and allows any background for back views and model views (Zalando image guidelines, updated 1 October 2026, checked 18 October 2026). Shoot it in even light, at a resolution high enough to zoom.
2. Back view
Shoppers want to see fastenings, back prints, pleats and the rear silhouette before they buy. Shoot the back with the same framing and light as the front, so the two read as a pair. Keep this photo even if you only publish the front today: it is also the source that lets AI tools rebuild a garment's interior and generate back views later.
3. Side view
A side view shows length and depth — how far a coat falls, how full a skirt is, how a jacket sits at the hip. It is often skipped because it needs another setup. From a front photo, and a back photo when available, some AI tools can reconstruct a side view: Photta's ghost mannequin tool documents a side angle for exactly this (Photta API docs).
4. Fabric close-up
A close-up of the weave, knit or print answers the questions a shopper would answer by touch: is it thick, sheer, soft, structured? Baymard found that 25% of ecommerce sites do not provide sufficient image resolution or zoom, and that low resolution led to product abandonment in its testing (Baymard, 2 April 2020). A dedicated close-up is more reliable than hoping shoppers zoom. Shoot it in raking light so the texture shows, with the true colour.
5. Construction close-up
Seams, lining, pockets, buttons, zips and labels show quality. One or two close-ups of the details that justify the price do more than a line of copy. If you have a sharp full-garment photo, AI can render a close-up of a named area from it: Photta's detail photo tool takes a label such as "collar stitching" or "cuff detail" and a framing from extreme close-up to medium close-up, and a free analysis step suggests which details are worth showing (Photta API docs). Check every generated close-up against the real garment.
6. On-model view
Baymard recommends showing apparel on a human model, because shoppers use those images to judge fit, length and proportions, and it advises stating the model's measurements and the size they are wearing (Baymard, 1 December 2020). Add that line under every on-model image; it is the cheapest fit information you can give.
AI can put a garment on a model from a flat lay, hanger or mannequin photo, which makes on-model images possible for products that never reach a studio. Photta's apparel try-on, for example, takes the garment, a chosen model and a pose, and an optional back image of the garment (Photta API docs). One caution: channels set their own rules for generated images. Zalando accepts AI-generated images that meet its editorial and technical criteria and can reject any that break believability or misrepresent the product; it does not allow visible AI labels or watermarks, and by December 2026 AI-generated files must carry identifying data in their metadata or as an invisible watermark (Zalando image guidelines, updated 1 October 2026, checked 18 October 2026). Use AI model images where the channel allows them, and check each channel's rules first.
7. Ghost mannequin
A ghost mannequin image shows the garment's shape and the inside of the neckline without a model, and keeps a whole category consistent. Traditionally it is a composite of two or more photos assembled by a retoucher (Pixelz, updated September 2025); AI tools can generate it from one mannequin or flat photo. Our AI ghost mannequin workflow walks through the steps and the checks.
8. Flat lay
A flat lay shows the whole garment at once, from above, and suits basics, children's clothing, sets and social media. Lay the garment naturally without stretching it, straighten seams and hems, and shoot from directly above so it is not distorted. A flat lay does not show fit, so pair it with an on-model or ghost mannequin image.
9. Colour variants
Photograph every colour you sell, under the same light, with a colour reference in a test frame. Shoppers compare colours across images, and a variant that looks different on screen from the product they receive is one of the most avoidable reasons for disappointment. If you use AI to show colourways, compare each result with the physical sample before publishing.
10. In context
Baymard's research found that 42% of users try to gauge a product's scale and size from its images, and that 28% of the large sites it benchmarked provided no "in scale" image at all (Baymard, 30 May 2017). For clothing, that means at least one image where the garment is worn or styled in a realistic setting. Keep the product the subject: a busy scene that hides the garment answers no question.
How many images does a clothing product need?
There is no universal number, but channels set minimums. Zalando requires at least three compliant images for apparel, one of them a catalogue view (a model front crop or a primary packshot), and it rejects duplicate images (Zalando image guidelines, updated 1 October 2026, checked 18 October 2026). A practical set for your own store is the front view, the back view, one close-up, and one image that shows fit or scale — on a model, as a ghost mannequin, or in context.
Before you publish: a checklist
- The front view is uncropped, centred and framed like the rest of the range.
- A back view exists, even if it is not published yet.
- At least one close-up shows the fabric and one the construction.
- At least one image shows fit or scale, with the model's size stated if a model is used.
- Every colour variant is photographed or checked against the physical sample.
- Resolution allows zoom on the details that matter.
- Each channel's rules are met, including whether AI-generated models are allowed.
- Every AI-generated image has been compared with the real garment.
Frequently asked questions
What are the most important clothing product photos?
The front view, the back view, a close-up of the fabric or construction, and an image that shows fit or scale. Together they answer what it looks like, what is on the back, what it is made of, and how it fits.
Should clothing be photographed on a model or flat?
Both have a place. On-model images show fit and proportion, which Baymard recommends for apparel. Flat lays and ghost mannequin images show the whole garment consistently. Most stores use a model or ghost mannequin image as the main view and a flat lay or close-ups as supporting images.
Can I use AI to create clothing product photos?
Yes, for many images: on-model views from a flat or mannequin photo, ghost mannequin images, close-ups and side views. Check every output against the real garment, and check each channel's rules: Zalando, for example, accepts AI-generated images but can reject any that break believability or misrepresent the product.
How do I show the fit of a garment?
Show it on a model and state the model's height, measurements and the size they are wearing, as Baymard advises. A side view and a back view add length and cut. If you have a size chart, link it next to the images.
What resolution should clothing photos be?
High enough for shoppers to zoom into the fabric and details. Channels set their own minimums — Zalando's minimum is 762 × 1100 pixels, with 1801 × 2600 as best practice (Zalando image guidelines, updated 1 October 2026, checked 18 October 2026) — so export from a master that is larger than any channel requires.