Reorder point formula for components and raw materials
Reorder point = average daily use × average lead time + safety stock. Set it for a purchased part with variable lead times, scrap, open POs and the MOQ.
A reorder point tells you when to place the next purchase order so the part arrives before you run out. For a purchased component the formula is short, but you make four choices inside it: which use to count, which lead time to trust, how much buffer to add, and which stock to compare it to.
What is the reorder point formula?
Reorder point = d × LT + safety stock, where d is average daily use and LT is the average lead time in days. The first part covers what you expect to use while you wait, and safety stock covers the days that run worse than expected. Some ERPs call the same number the reorder level or order point.
Without safety stock, the order lands just as the last unit goes out, and any late or short delivery stops the build. You can drop it only when use and delivery barely move, such as a steady part from a local supplier who delivers the next day.
Keep the units consistent. If the lead time is in calendar days, daily use has to be use per calendar day, weekends included. If your plant runs five days a week and the supplier quotes in calendar days, divide weekly use by seven. Divide by five and your reorder point fires up to a week late.
The reorder point tells you when to order. The MOQ, pack size and order costs set how much, and you make that decision on its own.
How do you work out demand during lead time for a component?
For a component, daily use comes from the build schedule. Multiply each parent product's planned builds by the quantity per, add scrap, and sum across the parents that use the part. Multiply that total by the average lead time and you have the demand during lead time, the first term of the formula.
For example, a machined component goes into two pump housings and a service kit. The housings use 180 and 140 a day and the kit 80 a day, so the component sees 400 a day. With 2% scrap, gross that up to 408 a day. Over a 46.7-day average lead time, that comes to 19,040 units of lead-time demand. Leave out the scrap and you come up about 370 units short in each cycle.
If several parents share the part, use their combined daily use. Combine their variation too: when each parent's demand moves on its own, add the variances. The safety stock calculator does that sum for you.
Which lead time should the reorder point use?
Use the time from placing the PO to having the part in stock and ready to issue, measured on your own POs. Put the average of those actual lead times into the formula, and use the spread around that average for the safety stock. Treat the supplier's quoted number as a starting point until you have receipts.
Pull the order date, the promised date and the received date for at least ten POs from the same supplier. The promised lead time tells you what the supplier said. The actual lead time tells you what to plan on. If the standard deviation of the actual lead times is more than about a quarter of their average, treat the supply as unreliable and expect the lead-time term to dominate your buffer. If receipts land four days late on average, your average lead time runs four days longer than the quote. Put those four days into the lead time and keep them out of safety stock.
Add the wait before an order can go out. If you check stock once a week, or the supplier takes orders only on Mondays, a part can cross the reorder point the day after a check and wait up to seven days for its PO. Add that interval to the lead time. In the worked example below, a weekly check raises lead-time demand by 408 × 7 = 2,856 units, and safety stock rises a little too, because it now covers a longer window.
If you plug the worst case into the formula and add safety stock on top, you count the same risk twice. Use the average lead time in the first term and let safety stock cover the spread. Supplier lead time covers what a quoted lead time often leaves out.
How much safety stock goes into the reorder point?
When both use and lead time vary, safety stock = Z × √(LT × σd² + d² × σLT²). Z sets how often you accept dipping into the buffer: 1.645 covers 95% of replenishment cycles, a cycle service level. On most purchased components with lead times over a month, the lead-time term inside the square root dwarfs the demand term.
The safety stock calculator works this out from a supplier quote or your PO history and gives you the reorder point with it.
The max/average shortcut, (peak daily use × longest lead time) − (average daily use × average lead time), assumes your busiest day lands on the supplier's slowest delivery in each cycle. On long lead times it can ask for four times the buffer the part needs. The demand-only formula, Z × σd × √LT, errs the other way: it treats the lead time as fixed and leaves out most of the risk.
One component, five ways to set the reorder point
A worked example table for one illustrative purchased component used at 408 a day after scrap, with a supplier quote of 35 to 60 days. It compares the safety stock and the resulting reorder point under five methods: the combined formula gives 3,567 units and a reorder point of 22,607; lead time only gives 3,447 and 22,487; demand only gives 915 and 19,955; max/average gives 15,232 and 34,272; and planning on the 95th-percentile lead time gives a reorder point of 22,713 with no separate buffer. The combined formula and the percentile method land within about 100 units of each other. Demand only leaves out most of the risk, and max/average asks for about four times the buffer.
| Method | Safety stock | Reorder point | What it assumes |
|---|---|---|---|
| Combined: Z × √(LT × σd² + d² × σLT²) | 3,567 | 22,607 | Use and lead time both vary, with no link between them |
| Lead time only: Z × d × σLT | 3,447 | 22,487 | Use is steady; only the delivery date moves |
| Demand only: Z × σd × √LT | 915 | 19,955 | The supplier delivers on the same day each time |
| Max/average: (peak d × longest LT) − (d × average LT) | 15,232 | 34,272 | Your busiest day lands on the slowest delivery in each cycle |
| 95th-percentile lead time: d × P95 LT | Built into the lead time | 22,713 | Plans on the slow delivery itself; useful when your lead times skew late |
On a long, variable lead time, demand only leaves the part short by more than 2,600 units, and max/average asks for over 11,000 units more than it needs.
If your actual lead times skew late, with a long tail the bell curve misses, take the 95th percentile of your measured lead times and set the reorder point to daily use × that lead time. In the example it lands within about 100 units of the combined formula, a sign your inputs hold together.
How do you choose the service level?
Pick one objective for the whole stock, such as the least inventory for a given service floor, and set one target for it. Then let each part's service level follow from that target. Cheap, steady, high-use parts end up higher, because each point costs little, and expensive, erratic, slow parts end up lower.
The target sits on a curve of inventory against service. You build it by adding stock where each dollar cuts the most expected shortage. Most stock sits inside that curve, so you can often hold the same service for less money. One service level for every part in an ABC class costs more than it needs to, so use classes for defaults and attention, and set per-part targets that hit each class's total.
These are our starting points, and one rule sets them: raise the level where an extra point of service costs few units, and lower it where it costs many. Going from 95% to 97% raises Z from 1.645 to 1.881, so every buffer grows by 14%. On a smooth part that is 14% of a small buffer; on a lumpy part it is 14% of a buffer that is already large. Adjust each level once you know what a shortage of the part costs:
| Class and demand pattern | Starting cycle service level | Why |
|---|---|---|
| A, smooth | 97% | High use, so a shortage stops the most output, and a steady part's buffer is small, so the extra 14% costs little |
| A, erratic or lumpy | 95% | The same two points would add 14% to a buffer that is already large and expensive |
| B, smooth | 95% | Middle value, the usual default |
| B, intermittent or lumpy | 93% | Z falls to 1.476, a buffer 10% smaller than at 95% |
| C, smooth | 95% | Cheap to hold, so each point of service costs little money |
| C, intermittent or lumpy | 90% | Low value and erratic use; Z of 1.282 cuts the buffer by 22% against 95%, and these parts often suit min/max or MRP better |
Each target also puts a price on a missed unit. From the newsvendor critical ratio, p = h × CSL ÷ (1 − CSL), where h is the cost of holding one unit for a year. A part that costs $2 a year to hold, at 95%, implies about $38 per unit short. If that price is far from what a shortage of the part would cost you, change the target. Treat the figure as an approximation.
For components, set service from the kit. One short part stops the build, so what matters is the chance that every part of the kit is there. Ten parts at 98% each, short independently of each other, give a complete kit about 82% of the time (0.98^10). To reach 95% for the kit, each of the ten needs about 99.5%.
Say which service figure you mean. Cycle service level is the chance of no stockout in a cycle; fill rate is the share of demand served from stock, and it depends on order size. The two can sit far apart on lumpy demand. Take an illustrative part over 20 cycles. In 19, demand is 100 units and the buffer covers it. In one, a work order pulls 4,000 units and stock covers only 1,000. That is 95% cycle service, 19 of 20 cycles without a stockout, yet you served 2,900 of 5,900 units, a fill rate of about 49%.
Why compare the reorder point to inventory position, not on-hand stock?
Inventory position is on-hand stock plus open purchase orders, minus stock you have already allocated to work orders. It is the stock you can count on. If you compare the reorder point to on-hand alone, you reorder while a PO is on the way, and you count stock that sits on the shelf but already belongs to a job.
Inventory position decides, not on-hand stock
A worked example table with three illustrative cases for the same component and a reorder point of 22,607 units. In the first, 24,000 on hand looks above the reorder point, but 2,000 are allocated and nothing is on order, so the inventory position of 22,000 says order now. In the second, 20,000 on hand looks below the reorder point, but a confirmed open PO of 5,000 brings the position to 25,000, so you wait. In the third, the position of 23,000 says wait, but the open PO of 12,000 has slipped to 35 days away while the usable stock lasts about 27 days, so the open PO needs to be expedited.
| Case (illustrative) | On hand | Open POs | Allocated | Inventory position | What to do |
|---|---|---|---|---|---|
| Stock on the shelf is already promised to a job | 24,000 | 0 | 2,000 | 22,000 | Order now: the position is below 22,607, although on-hand is above |
| A confirmed PO is on the way | 20,000 | 5,000, confirmed for day 20 | 0 | 25,000 | Wait: the position is above 22,607, although on-hand is below. Ordering now buys the same parts twice |
| The open PO has slipped | 12,000 | 12,000, now due in 35 days | 1,000 | 23,000 | Expedite: the position looks safe, but the usable 11,000 lasts about 27 days at 408 a day |
Compare the reorder point to the stock you can count on, and check the dates on the open POs inside that total.
On a long lead time, you often have several POs open at once. With 408 a day of use and 46.7 days of lead time, an order of 5,000 covers about 12 days, so about four orders sit in transit at any moment. If you watch on-hand stock instead of inventory position, the reorder point fires again the day after you order, and you buy the same parts twice.
Your inventory position depends on the dates on the open POs. If the supplier has pushed a PO out three weeks without telling you, that PO still counts as stock on the way, and the position looks safe while the part runs short. Confirm the date on each open line before you trust the total; purchase order tracking covers how, and expediting covers what to do when a date has slipped.
How does MOQ change the reorder point?
The MOQ leaves the trigger where it is. It sets how much arrives, so it changes the stock you hold after each delivery. Order when the inventory position reaches the reorder point, then order at least the MOQ, rounded up to the supplier's pack size. The timing comes from the reorder point, and the quantity from the MOQ.
MOQ vs EOQ covers how to set the quantity when the supplier's minimum sits far above what the part needs.
A large MOQ means more stock on average and fewer orders. If one MOQ covers months of use, the part spends most of its time far above the reorder point, and the buffer comes into play only in the last weeks of each cycle. A small MOQ against a long lead time gives you the case above, with several orders open at once, and you have to track each one.
A large MOQ also changes which service measure to use. One big lot already covers most cycles, so a cycle service target overstocks the part; size its safety stock on fill rate instead.
Check the stock you expect when an order lands: the inventory position when you ordered, minus the use during the lead time, plus the order quantity. If that number exceeds what your shelf space, cash or shelf life allows, negotiate the MOQ and leave the reorder point alone.
Worked example: a purchased component
This is an illustrative example. Three parents use a component at 400 a day in total, and you scrap 2%, so you need 408 a day. The supplier quotes 35 to 60 days, usually 45. Read as a triangle, the quote gives a mean of 46.7 days and a standard deviation of 5.14 days. Daily use has a standard deviation of 81.5 after scrap. You want 95% of cycles to end without a stockout.
- Lead-time demand: 408 × 46.7 = 19,040 units
- Safety stock: 1.645 × √(46.7 × 81.5² + 408² × 5.14²) = 3,567 units
- Reorder point: 19,040 + 3,567 = 22,607 units
- Cash in the buffer at $3.20 a unit: $11,414
The lead-time term makes up 93% of the variance, so each day of spread your supplier removes takes buffer off each cycle. If the supplier delivered on the same day for each PO, your safety stock would fall to 915 units. For a part like this, confirmed and accurate delivery dates save you more stock than a better forecast. The safety stock calculator starts with this example, so you can change any input and watch the reorder point move.
Which method fits which part?
Sort parts in two steps. First a lifecycle gate: new, active, dead or non-stock. Then rank active parts by consumption value (ABC) crossed with their demand pattern, and move any part with a long or unreliable lead time to the formula with lead-time variation.
| Part | Method |
|---|---|
| New part (less than 8 weeks of use, or fewer than 5 days with any use) | Days of supply on planned demand |
| Dead (no issue in 12 months) | No replenishment |
| Smooth use, reliable supplier | Reorder point with EOQ and price breaks |
| Unreliable or long supply | Reorder point with demand and lead-time variation |
| Intermittent or lumpy use | Min/max from a discrete quantile |
| Supplier ships on a fixed schedule | Periodic review that orders up to a level, then rounds to the MOQ and multiples |
| MOQ large against lead-time demand variation | Safety stock sized on fill rate |
| Component with a long cumulative lead time | Reorder point with a kit availability floor |
Days of supply for every part looks simple, but it gives each part the same cover, so steady parts carry too much and erratic ones too little. Treat DDMRP buffers as an option to opt into rather than a default.
ABC is a first cut. Parts move between classes from quarter to quarter, and a rare, cheap part whose stockout stops a line still ranks as C. Rank your review queue by dollars at risk instead: excess value plus expected shortage exposure.
When is the reorder point formula wrong?
The formula assumes steady use and deliveries that don't affect each other. It breaks when you use the part in lumps for occasional work orders, when the production schedule already sets the demand, or when lead times run in streaks because the supplier falls behind on all your orders at once. Plan those parts through MRP.
Watch for these signs. If you use nothing for days and then pull the part for one large work order, the daily average describes no real day; plan those parts from the schedule. If you make a part to order for a single customer job, you know the demand; net it against the job instead of holding a buffer. If a supplier's late POs come in runs, or its lead time stretches when the whole market gets busy, size the spread on the bad stretch.
Two statistical traps sit inside the formula. For intermittent or lumpy parts, the normal Z × σ mis-stocks; take the buffer from a Poisson or negative binomial quantile. Two common cutoffs sort these parts: a part is intermittent when the average gap between demands passes 1.32 periods, and lumpy when the squared coefficient of variation of its demand sizes also passes 0.49. And scaling daily σ by √LT, with an average you estimated, leaves the buffer too small and service below target, most of all when the average rests on a few weeks of data. Measure the swing on weekly buckets, or on lead-time demand itself.
If a reorder point looks far too high, check the inputs before you blame the formula. Look for one PO that waited on a drawing change, use in working days against a lead time in calendar days, or two suppliers' receipts mixed into one history. When a supplier's new date means an open order will land after the part runs short, a readiness signal opens a task, and Cartograph hands it to your team or Scout Agent. Scout emails and calls the supplier to confirm the date or expedite the line, within the permissions your team sets.
Questions teams ask
Can I calculate a reorder point in Excel?
Yes. With daily use in B2, average lead time in C2 and safety stock in D2, the reorder point is =B2*C2+D2. For the safety stock, use =NORM.S.INV(0.95)*SQRT(C2*E2^2+B2^2*F2^2), where E2 holds the standard deviation of daily use and F2 the standard deviation of lead time.
Which lead time should I use: quoted, average or worst case?
The average of your actual lead times, measured from PO to stock ready to issue. Use the quote only until you have about ten receipts. The worst case belongs in safety stock, through the spread; put it in the first term as well and you count the same risk twice.
Do open purchase orders change my reorder point?
No, they change the stock you compare it with. Inventory position is on-hand stock plus open POs minus allocations. Order when the inventory position reaches the reorder point, and confirm the dates on the open POs before you trust it.
How does MOQ affect my reorder point?
It leaves the reorder point where it is. The MOQ sets how much arrives, so it raises your average stock and spaces orders further apart. Order when the inventory position reaches the reorder point, then order at least the MOQ, rounded up to the pack size.
What is the difference between a reorder point and a reorder level?
None. Reorder level, order point and reorder point all name the stock level that triggers the next order. Some ERPs also use a min/max pair, where the min plays the role of the reorder point.
How do I set a reorder point for a seasonal part?
Take daily use for the weeks the lead time covers, not the yearly average. When the build plan ramps up, use the planned builds over the next lead time, and recalculate before the season starts. If the schedule sets the demand, plan the part through MRP.
How often should I recalculate reorder points?
Once a quarter for most parts. Recalculate at once when a supplier quotes a new lead time, a parent product ramps up or ends, or your received dates drift from the promised ones. The lead time tends to move first, so keep your actual lead times current.
What is the difference between a reorder point and safety stock?
Safety stock is the buffer for bad days. The reorder point is the trigger, and it includes the safety stock plus the use you expect during the lead time. If use and delivery both run to average, the order lands as your stock reaches the safety stock, and you don't touch the buffer.
How do I set a reorder point for a component used in several products?
Sum the component's daily use across its parents, after the quantity per and scrap, and use that total as daily use. To combine the variation, add the parents' variances when each parent's demand moves on its own, or their standard deviations when they rise and fall together.
When should I use MRP instead of a reorder point?
Use MRP when the production schedule already sets the demand for the part. MRP orders against known requirements and dates, while a reorder point waits for stock to fall. Reorder points suit steady, high-use parts such as fasteners, packaging and common raw material. Material requirements planning shows how MRP works out those orders week by week.