You have decided to buy 120 units of a new shirt. The harder question is how many of each size. Get the split wrong and you sell out of medium in a fortnight while extra small sits there until you discount it. This guide shows how to build a size curve from your own sales and apply it to a buy.
What a size curve is
A size curve is the share of sales each size takes within one style. It is the tool buyers use to split an order. The curve belongs to your customers, not to the industry, so build it from your own orders and not from a supplier's standard pack.
- Source: Your own order history for similar products
- Output: A percentage per size that adds up to 100
Build the curve from past sales
Take a comparable product, such as last season's version or the closest fit and cut you have sold, and count units sold per size. Convert each to a share of the total.
| Size | Units sold and share |
|---|---|
| XS | 10 units, 5% |
| S | 40 units, 20% |
| M | 70 units, 35% |
| L | 50 units, 25% |
| XL | 30 units, 15% |
Applied to a 120-unit buy, that gives 6, 24, 42, 30 and 18 units.
Correct for sizes that sold out
Sales only show what you had in stock to sell. If medium ran out halfway through the season, its share in the table is too low, and copying that curve repeats the stockout.
- Find the gap: Note the date each size sold out
- Compare fairly: Work out each size's share using only the weeks when every size was in stock
- Net out returns: Use units kept, not units shipped. A size that is often bought and sent back looks stronger than it is
Use a different curve for each fit and category
One curve for the whole store hides real differences. An oversized cut shifts sales towards smaller sizes. A slim cut shifts them up. Categories with many sizes behave differently again.
- By fit: Keep separate curves for slim, regular and relaxed cuts
- By category: Tops, bottoms and footwear each need their own
- By color: A core color can take the full size run. A fashion color may only justify the middle sizes
Handle minimums and edge sizes
Supplier minimums and packs rarely match your curve exactly. Round towards the sizes that sell fastest, because a stockout there costs you more than one spare unit.
- Pre-packs: If the supplier ships a fixed ratio, compare it with your curve and top up the gaps with open stock where you can
- Edge sizes: Decide on purpose whether to carry them. One or two units is a service to those customers. A deep buy is a risk
- Reorders: Reorder by size, not by style. Each size is its own SKU with its own rate of sale
For the reorder side, see how to calculate a reorder point and how to calculate safety stock.
Doing this at scale
Because each size is its own SKU, StockSense forecasts demand for each one from your order history, using velocity, trend, seasonality and lead-time-aware safety stock. It recommends a reorder point, quantity and date per SKU, ranks SKUs by revenue at risk, and drafts purchase orders grouped by supplier. See StockSense.