In this article
Do not add safety stock merely because a supplier “sometimes runs late.” Measure the distribution from purchase-order release to saleable stock, then separate demand uncertainty from lead-time uncertainty. A practical reorder point is expected demand during lead time plus a buffer that supports a chosen, fundable service level.
This matters to ecommerce operators, distributors and project suppliers in Qatar and the Gulf. Delay can occur in production, transport, customs or inspection. Measuring vessel arrival instead of usable-stock availability makes the lead time look better than the operation actually experiences.
Define lead time from buying decision to usable inventory
For every receipt, record purchase-order approval, supplier confirmation, factory departure, arrival, customs clearance, inspection and release for sale. Use the first and last dates for inventory policy; use the intermediate dates to diagnose causes. Do not confuse a supplier promise with an observed duration, or exclude late receipts as “abnormal.” Those observations are exactly what the decision must see.
NC State’s inventory tutorial describes safety stock as a cushion against forecast and lead-time uncertainty, while noting that more safety stock increases availability but also inventory and cost. There is no free service level.
Measure a distribution, not only an average
Start with at least 12 comparable receipts for the same SKU or product family when available. Calculate median, mean and the 90th-percentile lead time—the duration within which 90% of receipts arrived. If the mean is 18 days and the 90th percentile is 27, saying “the supplier takes 18 days” hides nine days of exposure in slower receipts.
Measure daily demand separately. Demand may be stable while lead time varies, or the reverse, or both may change in the same season. A simple formula is inadequate when demand spikes also overload the supplier; the risks are correlated and need a stress scenario.
Worked example: supplier variability and a buffer decision
Assume daily demand of 10 units, an 18-day mean lead time and a 27-day 90th percentile. If demand is stable and the pilot covers lead-time variation only, expected demand during the mean is 180 units. The lead-time buffer is 9 × 10 = 90 units, producing a simplified reorder point of 270 units.
At a landed cost of QAR 80 per unit, the buffer ties up QAR 7,200. Assuming a 15% annual holding rate, it costs about QAR 1,080 per year. Compare that with shortage exposure. If unit contribution is QAR 35 and the nine-day delay occurs in 20% of cycles, simplified expected lost contribution per cycle is 10 × 9 × 20% × 35 = QAR 630. Across six cycles, that is QAR 3,780 before expediting expense or service impact.
| Item | Assumption |
|---|---|
| Daily demand | 10 units |
| Mean lead time | 18 days |
| 90th-percentile lead time | 27 days |
| Lead-time buffer | 90 units |
| Capital in buffer | QAR 7,200 |
| Assumed annual holding cost | QAR 1,080 |
This is a teaching example, not an inventory recommendation. Backorders may preserve a sale. Perishable or fast-changing products may make excess stock more dangerous than shortage. Add minimum order quantities, storage, obsolescence and working-capital constraints to the decision.
Do not assign every SKU the same service level
Segment items by contribution, customer criticality, substitutability and recovery time. A high-contribution item with no substitute deserves a different target from a slow item that can be bought on demand. This analysis complements landed-cost and sourcing tests; it does not replace them.
Ask why lead time varies, not merely for a new average. The intervention might be a weekly order cut-off, reserved production capacity, a clearer request-for-quote specification, or a second source for part of the volume. Inventory absorbs the symptom; it does not repair an ambiguous buying process.
Measures and model limits
- Observed time from purchase order to saleable stock.
- 90th-percentile lead time and its direction, not only the mean.
- Availability rate and stockout days for critical SKUs.
- Stranded inventory value and expediting expense.
- Supplier promise accuracy and post-arrival defect rate.
Re-estimate after a change in route, factory, customs process or inspection terms. Do not apply a calm year to a different season, or combine suppliers in one distribution. A similar timing principle appears in rental-asset availability: quantity is insufficient when the item is not ready at the promised time.
Run a six-week decision pilot
- Select five critical SKUs with real sales history.
- Extract recent receipts and calculate mean and 90th-percentile lead time.
- Calculate a separate reorder point with explicit assumptions for each SKU.
- Apply the change to only two items and keep three as comparisons.
- Review availability, tied-up capital, expediting and write-offs before scaling.
The executive decision is not “hold more stock.” Choose a service level with a visible cost. If the buffer costs less than expected shortage and obsolescence risk is controlled, test it. If the root cause is poor specifications or an unreliable supplier, repair that cause instead of financing it with permanent inventory.
### Excerpt
Measure supplier lead-time variability and set a practical reorder point without tying up more working capital than needed.
Contact on WhatsApp
