Broken Size Curve

A broken size curve happens when one or more sizes of a product sells out while the rest remain available. The product still looks in stock on a listing page, but the sizes that actually sell are gone — so conversion rate falls even though traffic and inventory value look healthy.

Why it happens

Demand across sizes is not uniform. In most apparel categories the middle of the range — often M and L, or a 38–40 in footwear — accounts for a disproportionate share of units. If a buy is placed with an even spread across sizes, or the size profile is copied from a previous season with a different customer mix, the middle sells out first while the tails sit.

The result is a product that is technically 60% in stock and practically unsellable to most of the people who want it.

Why it is easy to miss

Product-level reporting hides it. At the parent level the item shows available inventory and, often, strong historical performance — so it keeps its placement in campaigns, feeds and merchandising. Meanwhile the conversion rate on that product is quietly collapsing, and paid traffic is still being sent to it.

The cost shows up in two places at once: lost sales on a product customers wanted, and ad spend against a page that can no longer convert. That is one of the more common sources of wasted ad spend in apparel.

How to measure it

Track size availability as a ratio rather than a binary in-stock flag: what share of the size range is available, and what share of historical demand does the available range represent. A product with 4 of 7 sizes in stock may still cover 80% of demand, or only 30%, depending on which four.

Pair that with sell-through rate per size to see the curve breaking before it breaks — the middle sizes accelerating away from the tails is the signal.

What to do about it

• Exclude at-risk products from ads automatically rather than manually, since the window between a curve breaking and the spend being wasted is days, not weeks.
• Reorder on the size profile, not the total. Replenishing the full range in even proportions recreates the problem.
• Feed real size demand back into buying. The curve that broke tells you what next season's buy should look like.
• Consider consolidation. Sometimes the right answer is to stop promoting the product and clear the remaining tail sizes rather than replenish.

Dema tracks size-level availability against demand and can exclude stockout-risk products from ad platforms automatically — see sell-through forecasting.