What actually happens when the shelf is empty
Everyone agrees stockouts cost sales. Far fewer people can say how much, or where the sale goes instead — and the answer is more useful than the agreement, because most of it is invisible in your own till.
How often the shelf is empty
The reference study is Gruen, Corsten and Bharadwaj's 2002 report for the Grocery Manufacturers of America, which drew on 52 separate studies — 16 previously published and 36 proprietary — and put the worldwide average out-of-stock rate at 8.3%.
Roughly one item in twelve, unavailable when a customer looked for it. In a fast-moving grocery category, on an ordinary day, with no crisis.
Where the sale goes
The more valuable contribution is the breakdown of what the shopper does next.
| Response | Share |
|---|---|
| Buys the item at another store | 31% |
| Substitutes a different brand | 26% |
| Substitutes the same brand, different size or form | 19% |
| Delays the purchase | 15% |
| Does not buy at all | 9% |
The 40% that walks out
Add the first row and the last: 31% go to a competitor and 9% buy nothing, so about 40% of stockouts take the sale out of your shop entirely. A further 15% merely postpone it, and may or may not come back.
The 45% who substitute are the reason retailers underestimate this. Those customers still pay you. The till rings, the day's total looks normal, and the only trace of the problem is a line that sold nothing and a customer who bought their second choice.
Which is the real insight here: a stockout is not a visible event in your sales data. It is an absence. You cannot find it by looking at what sold; you can only find it by knowing what you had.
Why the substitutions are not free either
Even the sales you keep cost something. A shopper pushed to a different brand may stay there. A shopper pushed to a larger size learns your shelf is unreliable. And a shopper who went to a competitor for one item did their whole basket there that day.
There is a later academic overview of this literature in Corsten and Gruen's work on shelf availability, if you want the research history rather than the single figure.
What to do with it
The practical consequence is about *when* you find out. Nothing in the breakdown above is fixable at the moment the shelf empties — by then the 40% has already happened. It is fixable only before, which means:
- A reorder point set from actual sell-through and the supplier's actual lead time, not from a round number.
- An alert that fires while there is still cover left to reorder within, not when the count reaches zero.
- A count you trust, since a reorder point computed from a wrong stock figure fires at the wrong time — which is why this post and the one on record inaccuracy are the same problem.
Sources
- Gruen, T., Corsten, D. & Bharadwaj, S. (2002). Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer ResponsesGrocery Manufacturers of America
- Corsten, D. & Gruen, T. On Shelf Availability: An Examination of the Extent, the Causes, and the Efforts to Address Retail Out-of-StocksSpringer
- DeHoratius, N. & Raman, A. (2008). Inventory Record Inaccuracy: An Empirical AnalysisManagement Science 54(4), 627–641
Figures on this site are attributed to the organisation that published them, with a link. Where sources disagree, we say so rather than choosing the flattering number.