The short answer
- A stockout is a customer leaving without the product they came for. In large grocery studies, nearly one in three then buys it at another store.
- Most stockouts start in the shop, not at the supplier: orders placed too late, stock records that are wrong, product in the back room but not on the shelf.
- If your supplier’s delivery times vary, your safety stock has to cover that too. It can be double what a sales-only calculation suggests.
- Recalculate reorder points before every busy period (Christmas, back to school, Black Friday) instead of keeping last month’s.
- A 20-minute weekly routine and a “couldn’t supply” log catch many problems before a shelf goes empty.
What running out actually costs
A widely cited source on this is a 2002 report by Thomas Gruen, Daniel Corsten and Sundar Bharadwaj for the grocery industry’s trade bodies. It pulled together 40 studies of empty shelves and surveyed more than 71,000 shoppers worldwide. These are supermarkets, not corner shops, and the data is old, so treat the numbers as orders of magnitude.
The rest either put off the purchase (15%) or don’t buy at all (9%). For a small shop, the headline is simple: almost a third of disappointed customers go to a competitor, and some of them stay there. Those who switch to a substitute tend to pick a cheaper one, the report adds, so even the “saved” sale earns less.
The average stockout rate across the studies was 8.3% of items, and noticeably higher for fast sellers and promoted lines. In one study the report cites, the fastest-moving 10% of items accounted for 45% of all stockouts. Your best sellers run out most, precisely because they sell fast.
Where stockouts come from
The same report splits the causes like this: about 47% come from store ordering and forecasting (ordered too late or too little), 25% from product that’s in the building but not on the shelf, and 28% from upstream (supplier, warehouse). Roughly three in four stockouts are decided inside the shop.
In a small business, that usually means one of five things:
- The order nobody placed. The item dropped below its reorder point three days ago and nobody noticed.
- Phantom stock. The records say 12, the shelf has 2. Breakage, theft or a typo: you think you have it, so you don’t reorder. In a study of nearly 370,000 inventory records across 37 stores of one retailer, DeHoratius and Raman (2008) found 65% were inaccurate.
- It’s out the back. Three boxes in the stockroom, an empty shelf since lunchtime.
- The late supplier. They say five days and sometimes take eight.
- The spike you didn’t plan for. A holiday, a promotion, a post that took off.
The first two are fixed with same-day stock records and regular counts, covered in our guide to inventory management for a small business, which also explains the basic reorder point. The other three need slightly better maths.
Safety stock when your supplier is unreliable
The standard reorder point is: average daily sales × lead time + safety stock. Many shops calculate safety stock from sales swings alone. That ignores half the risk: a supplier who delivers in four days one week and eight the next.
The Institute for Supply Management (ISM) gives a formula that covers both, when sales and lead times vary independently:
Safety stock = Z × √(average lead time × (std dev of daily sales)² + (average daily sales)² × (std dev of lead time)²)
Example: a coffee shop in Manchester sells an average of 8 bags of beans a day, with a standard deviation of 3 bags. The roaster delivers in 5 days on average, with a standard deviation of 1.5 days. The shop aims for 95% of order cycles without a stockout, so Z = 1.65.
- Sales variation only: 1.65 × 3 × √5 = 1.65 × 3 × 2.236 ≈ 11.1, rounded up to 12 bags.
- Sales and lead time together: 9 × 5 = 45; 64 × 2.25 = 144; √(45 + 144) = √189 ≈ 13.75; 1.65 × 13.75 ≈ 22.7, rounded up to 23 bags.
The late deliveries nearly double the safety stock. Reorder point: 8 × 5 + 23 = 63 bags. With the first calculation (40 + 12 = 52), a seven-day delivery at normal sales (56 bags) would already have emptied the shelf.
To get those standard deviations, you need a few weeks of daily sales and the real order and delivery dates. A spreadsheet does the rest with STDEV. Our inventory spreadsheet template shows the columns to set up.
Choose how much protection you want
Zero stockouts doesn’t exist, or costs a fortune in stock sitting on shelves. The better approach is to pick a service level for each type of product. Using the coffee example (√189 ≈ 13.75):
| Target service level | Z | Safety stock | Reorder point |
|---|---|---|---|
| 90% of cycles without a stockout | 1.28 | 1.28 × 13.75 ≈ 17.6 → 18 | 58 |
| 95% | 1.65 | 1.65 × 13.75 ≈ 22.7 → 23 | 63 |
| 99% | 2.33 | 2.33 × 13.75 ≈ 32.0 → 33 | 73 |
90% of cycles without a stockout
Z1.28
Safety stock1.28 × 13.75 ≈ 17.6 → 18
Reorder point58
95%
Z1.65
Safety stock1.65 × 13.75 ≈ 22.7 → 23
Reorder point63
99%
Z2.33
Safety stock2.33 × 13.75 ≈ 32.0 → 33
Reorder point73
Going from 95% to 99% means holding 10 more bags at all times. At £6 each wholesale, that’s £60 tied up in one product. Worth it for the house blend people come in specifically to buy. For a side line customers will happily swap, 90% is often enough. These percentages are business choices, not standards.
Plan for seasonal peaks
A reorder point built on October sales won’t survive December. Before each busy period, look at the same weeks last year and recalculate.
Continuing the example: in December the shop sells 12 bags a day instead of 8. Keeping the same standard deviations for simplicity (3 bags, 1.5 days):
- 9 × 5 = 45; 144 × 2.25 = 324; √(45 + 324) = √369 ≈ 19.21;
- safety stock: 1.65 × 19.21 ≈ 31.7, rounded up to 32;
- reorder point: 12 × 5 + 32 = 92 bags.
Keeping October’s trigger (63) in December means ordering 29 bags late, nearly two and a half days of December sales. In practice, sales usually get more volatile during the holidays too, so if you have last year’s figures, use them. And warn your supplier early: lead times can stretch when all their customers order at once.
The same logic applies to any demand you create yourself: a promotion, a product featured on Instagram, back-to-school week at a stationery shop. Raise the trigger before the spike, not during it.
The weekly routine
Most stockouts are visible a few days in advance. You just have to look at the right moment.
- Go through products below their reorder point and place the orders
- Physically count your 5 to 10 best sellers and correct the records
- Check best sellers’ shelves are full, and move stock from the back room
- Read this week’s “couldn’t supply” log
- Note every late delivery and by how many days
- Look at the next four weeks: holiday, promotion, school term?
The “couldn’t supply” log is the most underrated tool here. A lost sale shows up nowhere in your till data: the customer just walked out. Write down every request you couldn’t meet, even for products you don’t stock. After a month you’ll know which stockouts really cost you, and what your customers are asking for.
When you run out anyway
- Give a real date. “Back on Thursday” beats “soon”.
- Offer a substitute. The 2002 research suggests nearly half of shoppers will take another product. Point them to one with a decent margin.
- Take their number and message them when it arrives. That’s a delayed sale instead of a lost one.
- Update your online listings, or show the restock date. Selling what you don’t have leads to cancellations and lost trust.
- Find the cause the same day: missed order, wrong record, supplier, spike? Adjust that product’s reorder point or service level.
Out of house blend 250 g, 3 customers asked for it today. Remind me to call the roaster tomorrow at 9.
Ready in your “Shop” assistant: a note “House blend 250 g, out of stock, 3 unmet requests on 25/09” and a reminder “Call the roaster, house blend” tomorrow at 9:00. Save them?
Nothing is saved until you confirm.
Try Binome360 for freeSay a delivery, a sale or a breakage; your assistant prepares the stock movement and you confirm.
- One record per product, with cost price and selling price
- Stock in and out, typed or spoken, the same day
- Margins per product, so you know which stockouts cost most
- A recurring reminder for your Monday routine
Frequently asked questions
What is a stockout?
A stockout is when a product isn’t available to buy while a customer wants it: zero stock, or stock sitting in the back room instead of on the shelf. Its opposite is overstock, which ties up cash in products that don’t move.
How do you calculate your stockout rate?
Divide the number of items out of stock by the number of items you track, at a given moment. If 6 products out of 150 are missing on Monday morning, your rate is 6 ÷ 150 = 4%. Track it weekly, especially for best sellers.
How much safety stock should a small shop keep?
It depends on how much your sales and your supplier’s delivery times vary, and on the service level you choose. Calculate it per product with the formula above rather than holding a blanket “just in case” buffer on everything.
How do you avoid stockouts without overstocking?
Set protection product by product: high for best sellers, lower for items customers will happily swap. A calculated safety stock is usually cheaper than an across-the-board buffer.
What should you tell a customer when something is out of stock?
The expected restock date, a substitute if you have one, and an offer to let them know when it’s back. A specific date and a saved contact turn a lost sale into a delayed one.
In short
Stockouts are mostly decided inside the shop: a reorder point set too low, a wrong record, a peak nobody planned for. Safety stock that allows for late deliveries, reorder points recalculated before busy periods and twenty minutes every Monday make them rare. First step: start logging every request you can’t meet today, and read the list next Monday.
Sources
- Thomas W. Gruen, Daniel S. Corsten and Sundar Bharadwaj, Retail Out-of-Stocks: A Worldwide Examination of Extent, Causes and Consumer Responses, Grocery Manufacturers of America, Food Marketing Institute and CIES, 2002 (8.3% average stockout rate, shopper responses, causes, about 4% of sales lost): supplychain247.com (PDF).
- Nicole DeHoratius and Ananth Raman, “Inventory Record Inaccuracy: An Empirical Analysis”, Management Science, 54(4), 2008, pp. 627–641 (nearly 370,000 records, 37 stores, 65% inaccurate): doi.org/10.1287/mnsc.1070.0789.
- Institute for Supply Management (ISM), “Mastering Safety Stock Calculations: A Step-by-Step Guide”, 25 November 2025 (combined sales and lead-time formula, Z values for 90%, 95% and 99%): ism.ws.
Also available in Français.