Mon – Sat: 9:00 AM – 7:00 PM IST
Warehouse

Safety stock and reorder point

Two settings that quietly determine both your stockout rate and a large part of your working capital.

In short

Reorder point is the stock level at which a replenishment is triggered: expected demand over the lead time, plus safety stock. Safety stock is the buffer held against variability in that demand and lead time. Set them too low and you stock out; too high and you finance inventory you do not need. Both should be calculated per item, not applied as a blanket rule.

Why It Matters

Why this matters in practice

Safety stock is usually set once, by feel, and never revisited. The result is the pattern almost every business recognises: chronic shortages of a handful of items alongside years of cover on others. Both symptoms have the same cause, which is a buffer that bears no relationship to the variability of the item it is protecting.

Reorder point ties replenishment to lead time rather than to a calendar
Safety stock sizes the buffer to actual demand variability, item by item
Service level is an explicit choice with a visible cost, not an accident
Recalculating periodically releases cash tied up in stale buffers
How It Works

How it works in practice

Measure demand during lead time

Average demand per period multiplied by lead time in those periods. This is the base requirement.

Measure the variability

The standard deviation of demand, and separately of lead time. Items with steady demand need far less buffer than their volume suggests.

Choose a service level

95% or 98% are common. The implied safety factor rises steeply above 98%, which is where working capital disappears for marginal benefit.

Compute safety stock

Safety factor multiplied by the combined variability over lead time. Most ERPs will do this if the history is there.

Set and review

Reorder point equals lead-time demand plus safety stock. Review quarterly, and after any lead-time change.

Worked example

A worked example: what a service level costs

An item sells about 40 units a week with a standard deviation of 12. The supplier quotes three weeks, and actual receipts vary by about half a week either side. Expected demand over the lead time is 120 units.

At a 90% service level the safety stock works out at roughly 29 units, so the reorder point is about 149. Raising the target to 95% takes safety stock to about 38; to 99% it takes about 53. The first five points of improvement cost nine units of permanent stock; the last four points cost fifteen. That curve is the whole economics of the decision, and it gets steeper the closer to certainty you push.

The second finding is more useful than the first. Lead time variability contributes more here than demand variability does, because it multiplies against the full weekly demand rather than against its deviation. Getting the supplier to deliver reliably in three weeks — not faster, just consistently — would cut safety stock further than any forecasting improvement. That is a supplier conversation, not a planning one, and it is the kind of thing this calculation is for.

Key terms

The vocabulary, briefly

Term What it means
Reorder point The stock level that triggers a replenishment order: expected demand over the lead time, plus safety stock.
Service level The probability you intend not to stock out, expressed as a percentage. It is a choice, and it is expensive at the top end.
Demand variability How much demand moves around its average, usually as a standard deviation. The main input to safety stock.
Lead time variability How much supplier lead time moves. Often a larger driver of safety stock than demand variability, and more often ignored.
Stockout Demand you could not meet from stock. Worth recording even when the customer waits, or the number will always look good.
What Goes Wrong

Common mistakes worth avoiding

One blanket safety stock rule

"Two weeks of everything" guarantees shortage on volatile items and excess on steady ones, simultaneously.

Ignoring lead-time variability

A supplier whose lead time swings between two and six weeks needs more buffer than one that is reliably four, at identical demand.

Chasing 100% service level

The cost curve goes near-vertical past about 98%. Decide deliberately which items justify it.

Never revisiting after a supplier change

A new supplier means a new lead time and a new variability. The old reorder point is now wrong.

What to measure

Numbers worth tracking

Metric What it tells you What distorts it
Achieved vs target service level Whether your safety stock is doing what it was sized to do. Measuring by line rather than by order; a 98% line rate can still mean many incomplete orders.
Stockout frequency and duration How often and for how long, which matter differently to customers. Only counting stockouts that lost a sale, which ignores the ones that cost goodwill.
Lead time variance by supplier Usually the largest lever on safety stock, and the one that is a negotiation rather than a calculation. Using quoted lead time instead of measured receipts.
Excess stock value The cost side of the trade-off, so service level is a decision rather than a default. Valuing at cost without ageing, which hides stock that is excess and obsolete.
FAQ

Questions we are asked about this

It depends on the consequence of a stockout. A component that halts production justifies a higher level than a slow-moving spare. Differentiating by ABC class is the usual approach.
Not to finished goods, but very much to purchased components. Make-to-order businesses still buy against lead time and still face supplier variability.
It depends on what a stockout actually costs, and that varies enormously by item. A component that stops a production line, or a pharmaceutical a patient needs, justifies 98 or 99 percent. A slow-moving accessory a customer will happily wait a week for does not, and holding it at 99 percent is money sitting in a bin. Setting one level across the whole catalogue is the most common mistake — it is simultaneously too expensive for the tail and too risky for the critical few.
Calculated, and recalculated, for anything that matters. A fixed quantity set years ago encodes the demand and lead times of whenever it was set, and it ages invisibly. Fixed quantities are defensible for very low-value items where the calculation costs more attention than the stock is worth, but they should be a deliberate exception rather than the default.
It breaks it, if you use an annual average. A reorder point calculated on average demand will be far too low entering a peak and far too high leaving one. Either recalculate on a rolling recent window so the point moves with the season, or use explicit seasonal profiles. The warning sign is a stockout pattern that repeats at the same time each year — that is not bad luck, it is an averaging artefact.
Yes, by attacking variability rather than the buffer. More reliable supplier lead times, shorter lead times, better demand visibility and pooling stock across locations all reduce the safety stock needed for the same service level. Cutting the buffer without changing any of those simply converts stock into stockouts, and usually gets reversed within a quarter.
Related

Where to read next

ABC analysis for inventory

Read more

What is MRP?

Read more

ERP with warehouse management

Read more

Need this applied to your own systems?

We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.