The reorder point formula, and what getting it wrong costs
The reorder point is the stock level at which you place the next order, and the formula is short: average daily units sold × lead time in days, plus safety stock. For the worked example below that is 12 a day × 34 days = 408 units before any buffer at all. The whole difficulty sits in the safety stock, where three defensible methods give 168, 510 and 888 units for the same product — a fivefold spread, costing between £237 and £1,250 a year to hold. This page works all three, then does the one sum that actually decides between them: what the buffer costs against what a stockout costs.
The formula, and where each number actually comes from
Reorder point = (average daily units sold × lead time in days) + safety stock
That is the whole thing. Its reputation for difficulty comes entirely from the fact that all three inputs are usually wrong, and each is wrong in a different way.
Average daily units sold. Take a period long enough to be representative and divide. Ninety days is a reasonable default for a product that sells steadily. The mistake is using the last thirty days when those thirty days contained a promotion, or using a full year when the product only launched in March. Take the number from your own sales export, and take the same number every time so the figures stay comparable.
Lead time in days. This is not the number the supplier quotes. It is the number of days from you pressing send on the order to the stock being pickable on your own shelf, and it includes everything: the supplier’s production time, transit, customs clearance, inbound delivery, and the afternoon it sat in a box before anyone counted it. Sellers routinely use the production quote, which is often half the real figure. Go back through your last four or five orders and measure it.
Safety stock. The buffer that covers the difference between an average day and a bad one. This is the number the next two sections are about, because it is the only one you actually choose.
Harbourline Supply, an invented example
Harbourline Supply is invented and every figure attached to it is illustrative; it exists so the arithmetic has something to bite on. They sell a stainless steel travel flask.
- Units sold in the last 90 days: 1,080, so 1,080 ÷ 90 = 12 a day.
- Lead time, measured across their last five orders: 18 days of production + 12 days in transit + 4 days of clearance and inbound = 34 days.
- Busiest single day in those 90: 27 units.
- Worst lead time they have actually experienced: 48 days.
The reorder point before any safety stock is 12 × 34 = 408 units. Put plainly: the moment stock falls to 408 flasks, the next order has to be placed, because the 408 is exactly what will sell while the replacement is in the air. Order at 400 and you run out on the last day. Order at 300 and you are out of stock for nine days before the container arrives.
That nine days has a price, and it is easy to put a number on it: 9 × 12 = 108 units not sold, at Harbourline’s gross profit of £5.60 a flask, is £605 of gross profit gone, from a single decision to leave the order a fortnight.
Sizing the safety stock three ways
Safety stock exists because two things vary: how fast the product sells, and how long the supplier takes. There are three common ways to size it, they are not alternatives so much as three different questions, and the arithmetic connects them exactly.
| Method | Covers | Working | Units |
|---|---|---|---|
| Lead time variability only | The supplier being late, at normal demand | (48 − 34) × 12 | 168 |
| Demand variability only | A run on the product, at normal lead time | (27 − 12) × 34 | 510 |
| Both at once | Peak demand across the worst lead time | (27 × 48) − (12 × 34) | 888 |
The three are related, and seeing how makes the choice much easier to reason about. The worst-case figure of 888 is exactly the other two added together plus an interaction term:
168 + 510 + (27 − 12) × (48 − 34) = 168 + 510 + 210 = 888
That last 210 units is the part of the buffer that only exists if the busiest possible demand happens to land inside the longest possible lead time. It is the most expensive 210 units in the warehouse and the least likely to be needed, and it is where most of the argument about safety stock actually lives.
So the three reorder points are:
- Lead time buffer only: 408 + 168 = 576 units
- Demand buffer only: 408 + 510 = 918 units
- Both: 408 + 888 = 1,296 units
At 12 units a day, 1,296 units is 108 days of stock sitting on the shelf at the moment you place an order. That is a genuinely defensible number for a product you cannot afford to be out of, and a genuinely absurd one for a product with a substitute one click away. The formula will not tell you which you have. The next section will.
A note on service levels
Textbook treatments size safety stock from a service level and a standard deviation, which gives a smoother answer than the max-minus-average methods above. The idea behind it is worth having even if you never do the statistics: you are choosing what share of demand you are willing to fail to meet, and covering more of it costs steeply more, because you are buying stock for events that keep getting rarer. The max-based methods here approximate a very high service level, which is why they come out large. If your data is thin — fewer than a hundred days of sales, or three previous orders — the max-based version is the more honest one, because a standard deviation calculated from four observations is a decimal place pretending to be a fact.
What the buffer costs against what a stockout costs
This is the sum that decides it, and it is short. Both sides get converted into pounds a year and then compared.
The cost of holding the buffer
Holding stock costs money in four ways: the cash tied up, the space, the insurance and shrinkage, and obsolescence. Build your own rate from your own figures; Harbourline uses an illustrative 22% a year, made of 9% cost of money, 6% storage, 2% insurance and shrinkage and 5% obsolescence.
Their flask has a landed cost of £6.40.
| Safety stock | Value held | Annual holding cost at 22% |
|---|---|---|
| 168 units | £1,075 | £237 |
| 510 units | £3,264 | £718 |
| 888 units | £5,683 | £1,250 |
So the difference between the cautious buffer and the paranoid one is £1,250 − £237 = £1,013 a year.
The cost of running out
Put the same event in the same units. If the buffer is too small and the supplier takes the worst lead time they have ever taken, Harbourline is out of stock for 48 − 34 = 14 days. At 12 units a day that is 168 units not sold, at £5.60 of gross profit each:
14 × 12 × £5.60 = £941 per stockout event
The decision
Now the two sides meet. The larger buffer costs £1,013 a year more and prevents a £941 event. Divide:
£1,013 ÷ £941 = 1.08 stockouts a year
If Harbourline would otherwise suffer more than roughly one stockout a year, the bigger buffer pays for itself. If late deliveries happen once every two or three years, it does not, and the £1,013 is being spent to insure against something that mostly does not happen.
That single number is what this whole page is for. It replaces an argument about caution with an argument about frequency, and frequency is something you can look up in your own order history rather than have a view about.
Two honest caveats on it. The stockout figure counts only the gross profit on the units not sold. It does not count the customer who buys elsewhere and does not come back, which is real, usually larger, and not measurable from a spreadsheet — so treat £941 as the floor of the stockout cost rather than the whole of it. And it assumes a stockout lasts exactly the lead time overrun; if you can expedite a partial shipment, the real event is shorter and cheaper.
How much to order is a different question
The reorder point says when. It says nothing about how much, and the two get run together constantly. Order quantity is decided by three things the reorder point never sees: how much cash you have, what a shipment costs to move, and how long you are willing to have money asleep.
Harbourline has £9,000 available. At £6.40 landed that is a ceiling of 1,406 units. At 12 a day, or about 365 units a month, 1,400 units is 3.84 months of cover, so roughly 3.13 orders a year.
The cost of that choice is the average stock it forces you to carry:
Average stock = safety stock + (order quantity ÷ 2)
With the 168-unit buffer and 1,400-unit orders that is 168 + 700 = 868 units, or £5,555 of cash permanently asleep, costing 22% × £5,555 = £1,222 a year to hold.
Halve the order quantity to 700 units and the average stock falls to 168 + 350 = 518 units, £3,315 of cash, costing £729 a year. That is a saving of £493 a year — but it doubles the number of shipments from 3.13 to 6.26 a year. If each shipment costs £340 in freight, clearance and the admin around it, the extra 3.13 shipments cost £1,064.
| Order quantity | Orders a year | Average stock | Holding cost | Shipment cost | Total |
|---|---|---|---|---|---|
| 700 units | 6.26 | 518 units | £729 | £2,128 | £2,857 |
| 1,400 units | 3.13 | 868 units | £1,222 | £1,064 | £2,286 |
Bigger orders win here by £571 a year, and they win because the shipment cost is large relative to the holding cost. Change either figure and the answer flips: if freight were £90 a shipment rather than £340, the smaller order would win. This is exactly the trade-off the classic economic order quantity formula formalises, and the reason to do it as a two-row table rather than a formula is that the two-row table is auditable by anyone and the formula is not.
One constraint that overrides all of it: the order has to be at least the reorder point’s worth of cover, or you will hit the reorder point again before the previous order arrives and end up with two shipments in the air and no idea what you actually own.
What the reorder point cannot tell you
The formula is a good servant and a poor master. It is built on an assumption that is false for a great many products, and it is worth knowing exactly which.
- It assumes demand is roughly level. An average of 12 a day is useless for a product that sells 4 a day for ten months and 60 a day in December. For seasonal lines, run the formula on the demand rate for the period the stock will actually cover, not on the annual average, and place the pre-season order against the peak rate.
- It assumes you have a lead time history. A new supplier has none. Until you have three or four completed orders, use the supplier’s quote plus a generous allowance and treat the first reorder point as provisional.
- It cannot see a shutdown. Factory holidays, port congestion and a supplier taking a large order from somebody else are all lead-time events that no history predicts. The only defences are asking and holding more.
- It does not price the lost customer. The stockout cost above is the gross profit on units not sold. The customer who tried once, found nothing and never returned is worth more and cannot be counted. Where that risk is high, hold more than the arithmetic says.
- It assumes one pool of stock. If the same product sells through three channels from one shelf, the reorder point has to run on total demand, and if the channels hold separate stock, each needs its own. Running one reorder point against one channel’s sales is the most common way to be out of stock while technically holding plenty.
- It cannot tell you whether to stock the product at all. A line that ties up cash for four months to earn a thin margin may be failing a test the reorder point never applies. That question is turnover, and it is worked through in inventory turnover and what dead stock costs.
The inputs this formula runs on also feed the two sums either side of it: what a unit really cost to get onto the shelf, in how to calculate landed cost per unit, and what it has to sell for, in how to price a product so it actually makes money.
Arithmetic and general information only — not financial, tax, legal or investment advice. Your own figures, and your own accountant, decide what any of this means for you.
Do the quick version free
free break-even calculator — it gives you the monthly unit volume your fixed costs need, which is the demand figure the reorder point formula runs on. It runs in your browser and nothing you type is sent anywhere.
See inside the toolkit first
The Business Sale Readiness Toolkit sample shows every sheet, every row of the inputs sheet and the actual Excel formulas — no email, no account. If the sample is not worth your afternoon, the full workbook will not be either.
The tool for this job
The Small Business CFO Operating System ($39) is a working spreadsheet that holds unit costs, monthly volumes and what each line leaves behind, so the demand and holding figures the reorder point needs come out of your own records rather than a guess. One-time purchase, instant download.
Frequently asked questions
What is the reorder point formula?
Reorder point equals average daily units sold multiplied by lead time in days, plus safety stock. In the worked example on this page that is 12 units a day times a 34 day lead time, which is 408 units before any buffer. The lead time is measured from placing the order to the stock being pickable on your own shelf, not the production time the supplier quotes.
How do I calculate safety stock?
There are three defensible methods and they answer different questions. For supplier lateness alone, multiply the worst lead time minus the average lead time by average daily sales. For a run on the product alone, multiply peak daily sales minus average daily sales by the average lead time. For both at once, multiply peak daily sales by the worst lead time and subtract average daily sales times average lead time. On the same product those give 168, 510 and 888 units.
How much safety stock is too much?
Compare what the extra buffer costs to hold against what a stockout costs, both per year. In the example, moving from a 168 unit buffer to an 888 unit one costs 1,013 pounds a year in holding at a 22 percent holding rate, and a stockout lasting the worst lead time overrun costs 941 pounds of gross profit. Dividing gives 1.08, so the bigger buffer pays only if stockouts would otherwise happen more than about once a year.
Is the reorder point the same as how much to order?
No, and running them together is a common and expensive mistake. The reorder point says when to place the order. How much to order is set by your available cash, by what a shipment costs to move, and by how long you are willing to have money tied up. In the worked example, ordering 1,400 units at a time costs 2,286 pounds a year in holding plus shipments against 2,857 pounds for 700 unit orders, because the shipment cost outweighs the holding cost.
What lead time should I use in the formula?
The measured time from pressing send on the order to the stock being pickable on your own shelf. That includes production, transit, customs clearance, inbound delivery and the time it sat in a box before anyone counted it. In the example that is 18 plus 12 plus 4, which is 34 days, against a production quote that would have been about half of it. Measure it across your last four or five orders rather than taking the supplier at their word.