Why the clothes you order online never fit
There is no such thing as a medium. Sizes are set independently by each brand, product photos are actively engineered to conceal fit, and the person modelling the garment was chosen because sample sizes fit them. Each of those is a different problem with a different fix.
The ordering-three-sizes habit is usually treated as a personal failing — you should know your size, you should read the chart. It is not. It is a rational response to a system with three independent sources of error, none of which you control.
Problem one: a size is not a measurement
Clothing sizes are not standardised in any binding way. Each brand defines its own size chart against its own fit model — a real person the brand's patterns are cut to — and then grades up and down from there. Two brands whose fit models have different proportions will produce two 'mediums' that are genuinely different garments.
This is why the effect is so much larger than people expect. It is not drift or sloppiness; the numbers were never referring to the same thing. And it compounds within a single retailer: H&M runs a dozen internal lines that share a label but not a pattern block, and Aritzia is six house brands behind one checkout, each with its own block.
The fix: ignore the size label and use the flat measurements. Measure a garment you already own that fits the way you want, and compare pit-to-pit and body length against the listing. You are then comparing two objects rather than two abstractions.
Problem two: the photo is engineered to hide the fit
A product image is not documentation. It is a sales asset, and the entire craft of e-commerce styling is making a garment look its best on a rail-thin model in a studio. That craft has specific, well-known techniques, and every one of them removes information you needed:
- Clipping. A bulldog clip at the back gathers excess fabric so the front reads as fitted. The garment you receive has that excess.
- Steaming and pinning. Hems pressed flat, collars pinned open, sleeves arranged. Nothing hangs like that after a wash.
- Pose. A hand on the hip, a turned shoulder or a step forward creates a waist that the cut does not.
- Lighting. Directional studio light sculpts a silhouette. Your bedroom mirror does not.
- Sample sizing. The photographed piece is usually a sample size, which is not necessarily the size shown in the dropdown.
None of this is dishonest exactly — it is the same craft that sells everything else. But the net effect is that the image is a claim about the garment's best possible presentation, and you are trying to read it as a claim about its typical one.
The fix: get an image of the garment on a body that is yours. A generated try-on has no styling assistant, no clips and no pose direction — it renders the piece falling on your proportions under coherent light. That is the entire reason the technology is interesting.
Problem three: the model is not built like you
Even a completely honest photograph of a garment on a 5'10" model tells you very little about the same garment on a 5'3" frame. Not because of size — because of proportion. Where a hem lands, where a waist seam sits, how far a dropped shoulder drops: all of these are ratios between the garment's geometry and yours.
The categories where this bites hardest are exactly the ones people return most:
| Category | What proportion decides |
|---|---|
| Midi dresses | Whether the hem hits mid-calf (elegant) or ankle (different dress) |
| Trousers | Where the break falls — and rise changes the whole silhouette |
| Blazers | Whether the button stance flatters or cuts you in half |
| Oversized shirts | Whether it reads as designed volume or as borrowed |
| Coats | Whether the hem hits at the knee or below the calf |
The fix: the only one that exists — see it on your own proportions. This is the problem a size chart is structurally incapable of solving, because the chart describes width and you have a length question.
Three problems, three fixes
| Problem | Fix |
|---|---|
| Sizes are not standardised | Compare flat measurements, not size labels |
| Photos are styled to flatter | Generate the garment on your own photo |
| The model's proportions are not yours | Same — see where it lands on you |
Notice that two of the three have the same fix, and it is not a size chart. Most people spend their effort on the sizing problem, which is the one that is easiest to reason about and the least likely to be what actually went wrong.
What this costs everyone
Returns are the visible cost, and they are not free for anyone. The buyer pays in time, printing labels and post office queues. The retailer pays in processing, and frequently in the item never being resold at full price. And a returned garment travels — often twice — for a decision that could have been made before it was ever picked from a shelf.
The interesting thing about moving that decision earlier is not that it makes returns marginally cheaper. It is that a large fraction of them stop being necessary at all, because the shape was wrong and you could see that in advance.
Move the decision to before the order.
Get the appFrequently asked
Why do clothing sizes differ between brands?
+
There is no binding standard. Each brand defines its sizes against its own fit model — a real person their patterns are cut to — and grades up and down from there, so two brands with differently-proportioned fit models produce genuinely different mediums.
Why do clothes look different in product photos?
+
Product photography is a sales asset, not documentation. Garments are clipped at the back to look fitted, steamed and pinned, posed to create a waist, and lit directionally to sculpt a silhouette. All of it is standard practice, and all of it removes information about how the piece actually hangs.
How do I stop returning so many clothes?
+
Separate the shape question from the size question. Use a try-on to see whether the silhouette works on your proportions, and compare your own measurements against the garment's flat measurements to pick the size. Most repeat returns are shape problems being treated as size problems.
See it on you, not on a model.
Utopia puts any of 400,000+ pieces on your own photo in about a minute. Free on the App Store.
Get Utopia