UBC

Business journal

Reorder point and safety stock for a small shop

How much safety stock does a small shop need? A worked example with weekly sales and lead times, a quick rule and a statistical method.

Author: UBC editorial team · Published

7 min read

Safety stock is the extra quantity you hold so that a busy week or a late delivery does not leave you with an empty shelf. The reorder point is the stock level at which you place the next order: average sales during the supplier’s lead time, plus the safety stock. This article shows two ways to work out safety stock, a quick one and a statistical one, with the same example shop so you can compare them.

Stock level line falling and being replenished in cycles, with a shaded safety-stock band at the bottom and a delivery truck
Original illustration drawn in code by the UBC editorial team; no third-party material. CC0 1.0. Not a photograph of UBC premises or equipment.

If you have not set up basic stock records yet, start with how to keep stock records in a small business. It explains the reorder point formula with a short example; this article goes one step further and asks how big the buffer should be.

Why a fixed buffer is not enough

“Always keep two weeks of stock” is a common rule, and it is wrong in both directions. It ties up cash in products that sell steadily and arrive on time, and it still leaves you short on products whose sales jump around or whose supplier is unreliable.

What decides the buffer you need:

  1. How much your sales vary from one period to the next, not just how much you sell on average.
  2. How long the supplier takes, and how much that varies.
  3. How costly a stock-out is. Missing a top seller is worse than missing a slow accessory.
  4. How costly the buffer is. Cash, shelf space and, for some goods, shelf life.

The data you need

For each product you manage this way, collect:

  • Units sold per week for the last 8 to 12 weeks, from your orders, not from your purchases. Remove weeks when you were out of stock, as sales then understate demand.
  • The actual days each of your last several deliveries took, from order date to goods on the shelf.

The example shop

You sell a water bottle. Over the last eight weeks you sold:

18, 22, 15, 25, 20, 17, 24, 19 units.

The average is 20 units per week. The spread around the average (standard deviation of the weekly figures) is about 3.5 units. If you do not know what a standard deviation is, a spreadsheet function (STDEV or DESVEST) calculates it.

Your last six deliveries took 12, 13, 14, 14, 15 and 16 days. The average lead time is 14 days (2 weeks), and the spread is about 1.4 days.

Cycle demand during the lead time is 20 × 2 = 40 units. The question is how much to add on top.

Method 1: the quick rule

Compare the worst case with the average case:

Safety stock = (highest weekly sales × longest lead time) − (average weekly sales × average lead time)

  • Highest weekly sales: 25. Longest lead time: 16 days, which is 16 ÷ 7 ≈ 2.3 weeks.
  • Worst case: 25 × 2.3 ≈ 57 units. Average case: 40 units.
  • Safety stock ≈ 17 units. Reorder point = 40 + 17 = 57 units.

This is easy to explain and to compute, but it protects against the worst two events happening together, so it is generous. It is a reasonable start for an expensive-to-miss product.

Method 2: the statistical rule

Choose how often you accept to be covered. A service level of 95% means you aim for no stock-out in roughly 19 out of 20 replenishment cycles. This level corresponds to a factor of about 1.65 (the z-value for 95%, found in any statistics table). Then:

Safety stock = z × √( lead time × (variation of weekly sales)² + (average weekly sales)² × (variation of lead time)² )

with the lead time and its variation expressed in weeks. In the example:

  • Lead time × (sales variation)² = 2 × 3.46² ≈ 24
  • (Average sales)² × (lead-time variation)² = 20² × 0.2² ≈ 16
  • Square root of (24 + 16) ≈ 6.3
  • Safety stock = 1.65 × 6.3 ≈ 10 units. Reorder point = 40 + 10 = 50 units.

So with the same data the quick rule says 57 and the statistical rule says 50. The difference is the price of the extra caution. Neither is “correct”; they are tools for making the trade-off visible.

Two simplifications you can make: if your supplier is very reliable, use only the first term (safety stock of about 8 units here); if your sales are very steady but the supplier is unreliable, the second term will dominate. This is the case where a better supplier or a shorter lead time cuts your buffer more than a better forecast.

You can check the reorder point itself with the free reorder point calculator: enter average daily sales (20 ÷ 7 ≈ 2.9), lead time in days (14), the safety stock and your current stock.

Adjust by product importance

You do not need the same service level for every product.

  • Top sellers and items that bring customers back: aim higher, around 97 to 99%. The z-value rises (about 1.9 to 2.3), the buffer grows.
  • Normal range: 90 to 95%.
  • Slow sellers and easily substituted accessories: accept more stock-outs, or order on demand.

Fix the service level per group of products, not per product. A simple A/B/C split by revenue is enough. If you already count items on a cycle by the same A/B/C groups, you have the structure ready.

Real-life complications

  • Supplier minimum order and case sizes. If the supplier ships in cartons of 24, you order 24, and the reorder point may have to rise so you do not wait for a half-empty carton.
  • Reservations and open orders. Compare the reorder point with stock you can still sell: on hand minus reserved, plus what is already on its way.
  • Seasonality. The average of an off-season is a poor guide to December. Use last year’s same weeks, and review the numbers before each season.
  • New products. With no history, start with a deliberately cautious buffer and replace it with real data after the first 6 to 8 weeks.
  • Selling on several channels. Safety stock is for the total demand across channels, but each channel might need its own published buffer so that one channel does not consume all the stock. See selling on several marketplaces.
Columns of weekly sales next to a shelf of boxes, with a dashed reorder level
Original illustration drawn in code by the UBC editorial team; no third-party material. CC0 1.0. Not a photograph of UBC premises or equipment.

What the buffer costs

More safety stock means less risk of a missed sale and more cash on the shelf. Put a number on it before deciding: multiply the units of safety stock by the purchase cost. In the example, 17 bottles at €8.50 cost about €145 of cash, against 10 bottles at €85. Compare that with the profit you would lose on missed sales; the article on real profit per product on marketplaces shows how to find that profit.

A monthly review in four steps

  1. Update weekly sales and lead times for your top products.
  2. Recalculate safety stock and reorder points.
  3. List products below their reorder point and products with no sales in the last 8 to 12 weeks.
  4. Write down every stock-out and its cause: forecast, supplier delay, booking error. Patterns in these notes tell you more than any formula.

A spreadsheet is enough to do all of this at first. If the number of products or channels grows and the sheet starts to strain, this comparison helps decide when to move to inventory software.

Useful tip. Do not recalculate everything at once. Start with the three products that run out most often and look after a month at how orders and stock-outs have changed.

Frequently asked questions

Which service level should I choose? Usually 90 to 95% for the normal range and higher for top sellers. Decide by the cost of a missed sale and the cost of the buffer.

What if I have no sales history? Start with a cautious buffer and recalculate after 6 to 8 weeks of real data.

Is a spreadsheet enough? To begin with, yes. When products and channels multiply, inventory software becomes worth it.

How UBC handles this See UBC capabilities →

← All articles