Stockout rate

How the stockout rate is worked out

There are two sums, and the gap between them is the whole point of this entry.

The first counts items: how many of the lines you intend to stock were at zero, divided by the number of lines you intend to stock. Take it as a snapshot on a fixed day, or as an average of daily snapshots, which is better and almost as easy. This version is cheap to produce and it measures your buying discipline.

The second counts demand: the order lines you could not serve, divided by the order lines you received. This is harder, because it requires your system to record demand it could not meet rather than quietly dropping it, but it is the version that measures what the stockouts cost.

Also decide what you are intending to stock. A line you deliberately order in only when a customer asks for it is a range decision, not a stockout, and leaving those in the denominator dilutes the figure until it stops being useful.

There is no good number

A builders' merchant with 40,000 lines, most of them slow, and a food distributor with 900 fast-moving lines are not measuring the same thing even when they use the same formula. And neither figure means much on its own, because the two sums can disagree in both directions.

So read them together, and read the item list rather than the percentage. The percentage tells you how many. The list tells you which, and that is the only part you can act on.

In practice

Those 112 empty lines out of 4,000 give a stockout rate of 2.8 per cent by item.

Weight them by demand and the picture changes. Over the same week the warehouse received 11,000 order lines, and the 112 empty items accounted for 90 of them. By demand, the stockout rate was 0.8 per cent. Most of the empties were slow movers nobody asked for, which is uncomfortable but not expensive.

Now take a different week. Only one item was ever at zero, a top-ten seller, and it was empty for two days. By item that is one in 4,000, around 0.03 per cent, and the report looks immaculate. By demand it was 400 of 11,000 lines, or 3.6 per cent, and it was the worst week of the quarter. Both weeks are real; a report showing only the item count would have flagged the wrong one. The figures here are an illustration.

What moves the figure

• Reorder points that match the real lead time. The commonest cause of a stockout is a reorder point set against a lead time the supplier last met two years ago. There is a reorder point calculator on this site.
• The service level you chose. A stockout rate above your target is sometimes a buying failure and sometimes exactly what you specified. See service level.
• Stock accuracy. Phantom stock reads as available and fails at the pick face, which is the most expensive version of this problem. See stock accuracy.
• Promotions. A campaign launched without a stock plan concentrates the entire quarter's stockouts into one fortnight on one item.
• Allocation between channels. When a webshop and a trade counter draw on the same pool, one of them goes empty first, and which one is a decision somebody should be making on purpose.
• Supplier variability rather than supplier speed. A delivery date that moves is harder to protect against than a long one that holds.