Reorder points
Replenishment triggers: lead-time demand, safety stock, inventory position, order quantity, review frequency and dated supply exceptions.
Tags: production, planning, inventory, replenishment
On this page
With 12 units/working day, a fixed 5-working-day lead time and a 20-unit buffer, the example triggers at an inventory position of 80 units or less. Usable on hand 55, outstanding eligible supply 40 and unfilled demand 10 give 85, so no new order is triggered. After 8 units are fulfilled, position is 77 units and crosses the threshold. The separate fixed lot rule in this example is 100 units. Check whether open receipts arrive before shortages.
Oracle’s reorder model adds expected lead-time demand to safety stock. Its quantity calculation is separate from the trigger. Oracle Inventory User's Guide, Release 12.2, Reorder Point Planning
Define item, location, stock eligibility and the supply and demand included in the position.
Distinguish the quantities
For the original examples below, define inventory position as usable on-hand quantity plus outstanding replenishment quantity minus unfilled demand already owed. This is an explicit analytical convention, not a claim that every ERP field uses this formula.
| Quantity | Meaning to document |
|---|---|
| Physical on hand | What is physically recorded at the location |
| Usable on hand | The portion eligible for the intended demand |
| Outstanding supply | Remaining replenishment, with dates and status |
| Unfilled demand | Obligations included in the chosen netting convention |
| Inventory position | The defined combination used for the trigger |
Specify treatment of reservations, backorders, inspection holds, expired lots, transfers and repair returns. Do not subtract an allocation again if it is already excluded from the usable balance, and do not count a purchase-order line's original quantity when part has already been received.
Software terminology requires its own interpretation. Oracle's documented reorder-point method compares on-hand plus planned receipts with safety stock plus forecast lead-time demand. Its report also has supply and demand cutoff dates. This product-specific definition must not be silently replaced with a differently netted “available” field. Oracle, planning and report parameters
Estimate demand over the right lead time
Under an assumed constant average demand rate d and fixed replenishment lead time L, expected lead-time demand is d × L. Use compatible time units and calendars. Five working days of demand are not necessarily five calendar days of supplier lead time.
Include the relevant delay from replenishment decision to usable stock: approval, purchasing or internal release, supplier/manufacturing time, transport, receiving and required inspection as applicable. Oracle explicitly includes preprocessing, processing and postprocessing in its order lead time. Oracle, order lead time
For seasonal, intermittent or project-driven consumption, a constant-rate average can obscure the actual protection requirement. Review known demand dates and supply dates, and determine whether forecast-based reorder points are appropriate or a time-phased planning method is needed. A component's average historical usage should not automatically override a committed production requirement.
Keep safety stock separate from normal demand
Safety stock is a chosen allowance for uncertainty beyond the expected lead-time demand. Specify the uncertainty being addressed, data period, target service measure and review method. Microsoft describes the reorder point as anticipating lead-time demand and safety stock as covering fluctuations toward a targeted service level. Microsoft, Handling reordering policies
Do not label an arbitrary percentage or a fixed number of days as a statistically established service level. The probability of avoiding a stockout, the fraction of demand filled immediately and the percentage of orders delivered on time are different measures. A buffer cannot guarantee performance when its demand or lead-time assumptions fail.
Keep the expected-demand estimate and buffer visible separately so changes can be explained. Increasing safety stock to compensate for consistently understated lead time can hide the input error and increase inventory without resolving the underlying supply problem.
Original continuous-review example
Assume one item at one location, demand averaging 12 units per working day, a fixed 5-working-day replenishment lead time and an explicitly chosen 20-unit buffer. These are invented planning assumptions, not a recommended service policy.
Reorder point = 12 × 5 + 20 = 80 units
At a review instant, usable on hand is 55, outstanding eligible replenishment is 40, and already owed unfilled demand is 10. Using the defined convention:
Inventory position = 55 + 40 − 10 = 85 units
With a hypothetical trigger at position ≤ 80, no new order is triggered at 85. After 8 additional units of demand are fulfilled, usable on hand is 47 and position is 47 + 40 − 10 = 77, triggering consideration of replenishment. If the separate fixed-order rule is 100 units, placing that order raises position to 177, while physical stock remains 47 until a receipt occurs.
Receiving the existing 40-unit order transfers 40 from outstanding supply to on hand; it does not by itself increase position again. Under steady expected demand, starting at exactly 80 and consuming 60 units during the five-day lead time would leave the 20-unit buffer. That illustration assumes no other obligations, delays or losses; it is not a stockout guarantee.
A scalar position also hides timing. If the existing 40-unit receipt is too late to cover an earlier shortage, the planner must address that dated shortage even though total position looks sufficient.
Periodic review changes the protection interval
Continuous review evaluates the trigger as relevant inventory events occur. A periodic-review policy checks at scheduled intervals and may order up to a target rather than place a fixed quantity whenever a threshold is crossed. NC State's tutorial explicitly distinguishes these systems and describes regular reviews with replenishment toward a restocking level. NC State Supply Chain Resource Cooperative, Periodic Review System, January 28, 2011
As an original deterministic comparison, if reviews occur every 3 working days and lead time remains 5 working days, demand just after a review may need coverage until the next review's replenishment arrives: 3 + 5 = 8 working days. At 12 units/day, expected protection-period demand is 96 units, rather than 60. With the same illustrative 20-unit allowance, an order-up-to target would be 116. A real uncertainty allowance must be assessed for the longer interval; copying 20 is only a numerical comparison.
Business Central uses time buckets to accumulate demand and evaluate projected inventory. Its fixed-reorder policy also checks existing supply within lead time before suggesting another order. Those are implementation details to verify, not proof that a nominal reorder-point policy is continuously reviewed. Microsoft, time buckets and existing supply
Release and maintain the policy
Document trigger equality, review frequency, order quantity or target level, minimum orders, pack multiples and authorized overrides. A threshold crossing produces a proposal under the chosen workflow; distinguish that proposal from an approved purchase, confirmed supply and received usable stock.
Reconcile inventory transactions and overdue orders before relying on automation. Monitor actual lead times, demand, stockouts, excess stock and obsolete balances. Reassess parameters after supplier changes, process changes or demand shifts rather than assuming the first calculation remains valid indefinitely.
Record item and location, eligible stock, netting, forecast, lead-time calendar, safety-stock rationale, trigger, quantity rule, review interval and exceptions. Keep effective dates.
Source versions
Reviewed 2026-11-02 PST. Oracle's cited collection is Release 12.2; no separately verified issue date is assigned to the chapter. Microsoft's page displays June 12, 2024. NC State's tutorial displays January 28, 2011. Examples and sensitivity calculations are original, bounded models, not company policy or software defaults.