Size curves

Everyone concentrates on how many units to make. Far fewer people think carefully about how those units split across sizes, and that split decides how much of the run you can actually sell.

Get it wrong and you end up in the worst possible position: sold out of the sizes people want, while holding stock of the ones they do not. You are out of stock and overstocked at the same time, and the leftover units are frequently the entire margin.

Why an even split is wrong

Size demand is not uniform. It clusters heavily in the middle and thins toward the ends, in essentially every adult audience.

Split a run evenly and the arithmetic is unforgiving. The middle sizes sell out early, at which point the drop is effectively over for most of the people still arriving — while the end sizes sit. You then discount those to clear them, which means the units that cost the most to hold are also the ones that earn the least.

The curve is not a refinement to apply once you are bigger. It matters most on a small run, because a small run has no depth to absorb the error.

The shape

Without your own data, work from the shape rather than any specific numbers:

• The middle two or three sizes carry the bulk of the volume.
• The size either side of the middle carries a meaningful share.
• The extremes are thin — real, worth stocking, but a small fraction.

That produces a curve that rises to a peak and falls away, not a flat line. Where the peak sits and how sharply it falls is what your audience determines.

What moves the peak

Fit is the biggest single factor. Oversized and boxy cuts pull demand down the range, because people buy a smaller size to get the intended silhouette. A fitted or slim cut pushes demand up, for the same reason in reverse. If you are selling the same artwork in two fits, you need two curves — using one is how a boxy run strands its larger sizes.

Audience. Skew by gender, age and region, all significantly. A curve built on one audience does not transfer to another, and this is where borrowing an industry-standard curve goes wrong.

Channel. Event and pop-up demand runs differently from online: buyers take what fits now rather than waiting, which flattens the curve. Online has more deliberate size selection and a sharper peak — and higher returns, which is its own inventory question.

Whether it is unisex. A unisex garment sold to a mixed audience produces a broader, flatter curve than the same garment sold to one group, because two overlapping distributions add together.

Getting a real answer

Preorder. The only method that tells you the truth rather than an estimate. People select their actual size, and you make what was actually asked for. It also validates the quantity, so it solves two problems at once — and it is why demand validation appears before inventory commitment in every sensible version of this process.

Your own history. If you have sold to this audience before, that data beats any general rule. Use sell-through by size, not units sold — the size that sold out first was under-ordered, and units sold alone cannot show you that.

Waitlist with size. Cheaper than a preorder and directionally useful. Asking for size at signup costs one field and tells you the shape.

Industry curves. A reasonable starting point when you genuinely have nothing, and worth treating as a prior to be replaced rather than an answer. They are averaged across audiences that are not yours.

Practical decisions

Do not cut the extremes to save money. Dropping the largest and smallest sizes looks efficient and reads badly — it tells part of your audience the product was not made for them. The cost of carrying a few units is usually smaller than the cost of that message.

Consider a shorter range instead of a thinner one. If the run genuinely cannot support the full range, offering fewer sizes honestly is better than offering all of them and being out of stock in half within a day.

Watch sell-through by size, not just total. Total sell-through hides the failure. A drop that sold 80% overall can be a drop that sold out of three sizes on the first day and never moved the rest, and those two outcomes need completely different decisions next time.

Record the curve you used and what happened. The second run is where this compounds, and only if someone wrote down what the first one taught you. That single habit is worth more than any starting curve.