A reorder point calculator answers a single operational question: at what stock level should a replenishment order be raised? Order too late and you sell out while the shipment is in transit. Order too early and capital sits on a shelf earning nothing. The reorder point is the level at which the remaining stock is just enough to cover expected demand during the lead time, plus a buffer sized to the uncertainty you face and the service level you have chosen.
Arb Digital publishes this in its free tools library beside the inventory turnover calculator, which measures how efficiently stock is being used, and the COGS calculator, which prices what is sitting there. This page is the trigger-level calculation, and it is the one that determines whether a product is available when someone tries to buy it.
What This Reorder Point Calculator Does
It computes the reorder point in two parts and shows both. Lead-time demand is the straightforward part: average daily demand multiplied by average lead time, which is what you would need if nothing ever varied. Safety stock is the buffer, sized from the combined variability of demand and lead time and scaled by the Z-score corresponding to your chosen service level.
Crucially it includes lead-time variability, which many simpler calculators omit. That omission is not a rounding error — for most businesses supplier unreliability contributes more to stockout risk than demand fluctuation does, and a safety stock figure that ignores it is systematically too small.
The tool also converts the answer into days of cover, shows how often you would be placing orders at your typical order quantity, and puts an annual holding cost on the safety stock so the service level decision has a price attached rather than being chosen by feel.
How to Use It
- Measure demand over the same period you will use. If lead time is in days, demand and its standard deviation must be daily. Converting a monthly standard deviation to a daily one is not a matter of dividing by thirty.
- Include everything in lead time. Supplier processing, production, transit, customs and inbound receiving all count. Stock that has arrived but is not yet sellable is not stock.
- Measure lead-time variability from your own purchase history. The supplier's quoted lead time is a target. What matters is the distribution of what actually happened over the last dozen orders.
- Choose a service level per product, not per business. A high-margin bestseller and a slow-moving accessory should not carry the same buffer.
- Recalculate when anything moves. Reorder points are commonly set once and left for years, which is how a business ends up with buffers sized for demand it no longer has.
The Formula / How It's Calculated
The reorder point is the sum of two terms:
Reorder point = (average daily demand × average lead time) + safety stock
Safety stock combines both sources of variability under a square root, because independent variances add rather than the standard deviations:
Safety stock = Z × √(lead time × σdemand² + demand² × σlead time²)
Z is the standard normal value for the target cycle service level — 1.645 at 95%, 2.326 at 99%.
Worked example, matching the values the page loads with. Average daily demand is 120 units with a standard deviation of 30; lead time averages 14 days with a standard deviation of 3. Lead-time demand is 120 × 14 = 1,680 units. Inside the square root, the demand term is 14 × 30² = 14 × 900 = 12,600 and the lead-time term is 120² × 3² = 14,400 × 9 = 129,600, totalling 142,200. The square root of that is 377.1, so at a 95% service level the safety stock is 1.645 × 377.1 = 620 units. The reorder point is 1,680 + 620 = 2,301 units, rounding up. That is 19.2 days of cover, of which 5.2 days is buffer. Note that the lead-time term contributes 129,600 of the 142,200 total — 91% of the variance comes from the supplier, not from customers. MIT OpenCourseWare's Introduction to Operations Management covers the underlying inventory models.
Lead-Time Variability Usually Dominates
This is the finding that changes what a business should actually do about stockouts, and the worked example makes it concrete: 91% of the variance came from the lead time.
The reason is structural. The demand term is multiplied by lead time, but the lead-time term is multiplied by the square of daily demand. For any product moving in reasonable volume, that squared term grows fast. At 120 units a day, one extra day of lead-time uncertainty carries far more weight than a substantial swing in daily orders.
The practical implication is a reordering of priorities. Investing in a better demand forecast when lead-time variability dominates buys very little — you would be refining the small term. Reducing the standard deviation of lead time from 3 days to 1 day in the example cuts safety stock from 620 units to 270, a 56% reduction, without touching the forecast at all.
That reduction is bought through supplier management rather than analytics: fixed shipping schedules instead of ad-hoc dispatch, a second qualified supplier, tighter booking of freight, or holding a domestic buffer for an overseas line. It is worth pricing. At $3.50 of annual holding cost per unit, those 350 units of avoided safety stock are worth about $1,225 a year on this product alone, and the same intervention usually applies across every line from that supplier.
What a Service Level Actually Costs
Service level is where most of the money is decided, and it is usually chosen by instinct rather than arithmetic. The relationship is not linear, and the last few percentage points are extraordinarily expensive.
In the worked example, moving from 90% to 95% raises safety stock from 483 to 620 units. Moving from 95% to 99% raises it to 877. Moving from 99% to 99.9% raises it to 1,165. Each step buys less protection than the previous one and costs more stock than the previous one, because the tail of a normal distribution stretches out indefinitely. Reaching 100% is impossible at any price, which is worth saying plainly to anyone who asks for it.
There is also a distinction the term hides. Cycle service level — what this calculator uses — is the probability of not stocking out during a replenishment cycle. It says nothing about how badly you stock out. A different measure, the fill rate, describes the proportion of demand met from stock, and it is usually much higher than the cycle service level for the same buffer. A product at a 95% cycle service level frequently runs a fill rate above 99%, because the cycles that do stock out typically miss only a handful of units at the very end. If a customer-facing commitment is being made, the fill rate is normally the relevant number.
The right service level therefore varies by product and comes from comparing two costs: the holding cost of the buffer against the margin lost on an unmet sale, plus whatever a stockout does to customer behaviour. Fast-moving, high-margin items with substitutes readily available elsewhere justify high levels. Slow, low-margin items with patient customers do not.
Where the Model's Assumptions Break
The formula assumes normally distributed demand, independent variability in demand and lead time, and a stable average. Each assumption fails somewhere predictable.
Slow movers. A product selling two units a week is not normally distributed — its demand is a count with a long right tail, better handled by a Poisson model. Applying the normal formula to intermittent demand produces a buffer that is too small most of the time and pointless the rest.
Seasonality and trend. A single average daily demand figure assumes a flat baseline. If demand quadruples in the fourth quarter, a reorder point built on the annual average is far too low exactly when a stockout costs most. The correct approach is to recalculate for each season using demand from the comparable period, and the revenue forecast calculator is a reasonable starting point for projecting the level.
Correlated failures. The formula assumes demand and lead time move independently. In practice they often spike together — a demand surge across a supplier's customer base is exactly when that supplier's lead time extends. When that correlation exists, the calculated safety stock understates the requirement.
Stock that exists but cannot be sold. Damaged, reserved, mis-located or unreceived units count in the system and not in reality. A reorder point compared against an inaccurate on-hand figure produces the right answer to the wrong question, which is why cycle counting sits underneath any of this working. Inventory measurement itself is governed by accounting standards — IAS 2 Inventories sets out how stock is recognised and measured.
Arb Digital builds the demand side — search visibility and product pages that bring predictable, forecastable traffic to the lines you stock.
See Web Growth Services Talk to Arb DigitalCommon Mistakes to Avoid
- Ignoring lead-time variability — it frequently contributes the large majority of the total variance, so leaving it out makes safety stock far too small.
- Mixing time periods — daily demand with a weekly standard deviation, or a lead time in weeks against demand in days, produces answers that are wrong by whole multiples.
- Applying one service level to every product — the correct level depends on margin, substitutability and holding cost, and these differ enormously across a catalogue.
- Using the supplier's quoted lead time — what matters is the distribution of what actually happened on your last several orders, including the late ones.
- Setting reorder points once — demand levels, lead times and product mixes all drift, and a buffer sized for last year's conditions protects against last year's risks.
Related Free Tools From Arb Digital
Measure how hard your stock is working with the inventory turnover calculator, price it with the COGS calculator, and check the margin a stockout would cost with the profit margin calculator or the contribution margin calculator. The break-even units calculator sets the volume a line has to reach, the shipping cost per item calculator covers inbound and outbound freight, the revenue forecast calculator projects the demand this model needs, and the purchase order generator raises the order once the trigger is hit. Everything else is in the free online tools hub.
Frequently Asked Questions
The stock level at which a replenishment order should be raised. It equals the demand expected during the supplier lead time plus a safety stock buffer sized to the variability in demand and lead time and to the service level you are targeting.
By multiplying a Z-score for the target service level by the standard deviation of demand during lead time. That standard deviation combines both sources of uncertainty under a square root, because independent variances add rather than the standard deviations themselves.
Because it is multiplied by the square of average daily demand, while demand variability is multiplied only by lead time. For any product moving in volume the lead-time term dominates, which is why supplier reliability usually affects stockouts more than forecast accuracy does.
It depends on the product. Compare the annual holding cost of the extra buffer against the margin lost on unmet demand and the effect a stockout has on customer behaviour. High-margin fast movers justify high levels; slow, low-margin lines rarely do.
No. Cycle service level is the probability of not stocking out during a replenishment cycle. Fill rate is the proportion of demand met from stock, and it is normally much higher for the same buffer because a cycle that stocks out usually misses only a few units at the end.
They answer different questions. The reorder point determines when to order; the economic order quantity determines how much. Order quantity affects how often you reach the trigger, but it does not change the trigger level itself.
Not well. The formula assumes roughly normally distributed demand, and an item selling a couple of units a week has intermittent demand better modelled with a Poisson distribution. The normal approximation produces buffers that are unreliable at very low volumes.
Whenever demand levels, lead times or the product mix shift, and on a scheduled basis otherwise — quarterly is common. Seasonal lines need recalculating for each season using demand from the comparable period rather than an annual average.
