Size demand corrected for stock-outs
Restores the true size curve: divides the sales of each size by the days it was in stock. Shows which sizes are understated in the curve by sales, how many units were lost and how to split the purchase.
At a glance
- What question does it answer?
- What the true size curve is if some sizes were not in stock for the whole period, and how many sales were lost.
- What data is needed?
- Sales by size over the period.
- How to read the result?
- Compare the columns “by sales” and “true”: for sizes where the true share is higher, availability was limited. Lost units show the price of the shortage. Example on the default values: Size with the largest understated share: XL; Understatement of its share, pp: 10.6; Units lost to missing sizes: 65; Lost units to actual sales, %: 40.6%. The figures are illustrative: enter your own to see your result.
Why and how to reason
A curve by sales understates sizes that were out of stock. The rate of a size over its in-stock days shows the real demand. In the example sales are 10, 40, 60, 35, 15 with 60, 60, 60, 30 and 20 in-stock days: by sales L gives 21.9%, the true share is 31.1%, XL 9.4% and 20.0%; 65 units were lost, a purchase of 1,000 units is split 44, 178, 267, 311, 200.
When to use it
Apply before buying by size and when analyzing the season results, especially for large sizes where shortages are more frequent. The split by given shares is done by sizecurve, here the shares are restored.
How to read the result
Compare the columns of the share by sales and the true one: where the true one is higher availability was limited. Read sizes with a window of a few in-stock days carefully.
Common mistakes
The curve is built from sales at incomplete availability, in-stock days are taken by the warehouse and not by the store, sizes in stock for a few days are taken as they are.
What data is needed?
- Sales by size over the period.
- In-stock days for each size where it sells.
- The total purchase and the box multiple.
Inputs
| Field | Unit | Value in the example | Meaning |
|---|---|---|---|
| Total purchase, units | pcs | 1,000 | How many units of the model or group are split across sizes. The example value is illustrative. |
| Length of the period, days | days | 60 | Over how many days the sales and in-stock days are collected. In-stock days cannot exceed it. The example value is illustrative. |
| Minimum in-stock days for a reliable estimate | days | 7 | A size that was in stock for fewer days than this is flagged: a rate over a few days is random. The example value is illustrative. |
| Box multiple, units | pcs | 1 | The purchase is split in whole boxes; a remainder smaller than a box stays unallocated. 1 means no rounding. The example value is illustrative. |
How it is calculated
| Step | Formula |
|---|---|
| Size rate | v(s) = sales(s) ÷ in-stock days(s) |
| True share | d(s) = v(s) ÷ sum of v |
| Share by sales and shift | o(s) = sales(s) ÷ sum of sales; shift = d(s) − o(s) |
| Lost units | v(s) × (60 − in-stock days(s)) |
| Purchase of a size | total purchase × d(s); in whole boxes by the largest remainder |
Worked numbers
Default values from the app; in the interactive version you enter your own.
- Size with the largest understated share
- XL
- Understatement of its share, p.p.
- 10.6
- Units lost to missing sizes
- 65
- Lost units to actual sales, %
- 40.6 %
- Purchased by the corrected curve, units
- 1,000
- Unallocated because of the box, units
- 0
- Sizes with a short in-stock window
- 0
- By sales
- True
Data as a table
| Measure | By sales | True |
|---|---|---|
| XS | 6.3 | 4.4 |
| S | 25 | 17.8 |
| M | 37.5 | 26.7 |
| L | 21.9 | 31.1 |
| XL | 9.4 | 20 |
The size curve
| Size | Sales | In-stock days | Rate, units/day | Share by sales | True share | Shift, p.p. | Lost units | Purchase |
|---|---|---|---|---|---|---|---|---|
| XS | 10 | 60 | 0.17 | 6.3 % | 4.4 % | −1.8 | 0 | 44 |
| S | 40 | 60 | 0.67 | 25 % | 17.8 % | −7.2 | 0 | 178 |
| M | 60 | 60 | 1 | 37.5 % | 26.7 % | −10.8 | 0 | 267 |
| L | 35 | 30 | 1.17 | 21.9 % | 31.1 % | 9.2 | 35 | 311 |
| XL | 15 | 20 | 0.75 | 9.4 % | 20 % | 10.6 | 30 | 200 |
The curve by sales understates sizes L, XL: they were not in stock the whole period. A purchase by observed shares would leave them without stock.
What to do with the result: buy by the true curve, not by sales; the lost units show how much demand was lost because sizes were missing and justify a wider size run.
A typical mistake: building the curve from sales when the size was in stock for half the period and counting in-stock days by the warehouse, not by the store.
Nuances and limits
- Building the curve from sales with incomplete availability.
- Taking in-stock days by the warehouse and not by the store.
- Overlooking sizes with a few in-stock days.