Measuring Supplier Lead Time Until Stock Is Usable

Supplier lead time should be measured to the point where the received material can serve the intended production requirement. A gate arrival, goods receipt and quality release can occur at different times. If the replenishment model stops at arrival while production must wait for approval, the recorded lead time omits part of the actual protection period.

This guide helps procurement and inventory analysts create a timestamp record for lead-time variation. It concentrates on defining observations, handling partial receipts and separating supplier performance from internal delays. A buffer calculation should follow that work, rather than hiding inconsistent source data behind a standard deviation.

Define a beginning and an end

Write the operational question before selecting timestamps. To measure supplier execution, the start may be the supplier's accepted purchase order and the end the agreed delivery milestone. To measure total replenishment exposure, the start may be the internal replenishment signal and the end stock available for use. Both are useful, but they measure different processes.

Oracle's replenishment documentation identifies preprocessing, processing, transit and postprocessing components in buy-order lead time. Use that distinction as a prompt to examine the full path. Do not assume the fields or calculation in one ERP are the exact implementation of another system.

For the pilot record here, define total operational lead time from approved replenishment release to quality-released stock availability. Keep supplier acknowledgement, dispatch, arrival, receipt posting and release as separate milestones. If the starting approval itself takes time after the original signal, retain that interval separately and decide explicitly whether the planning model must cover it.

Use a stable observation key

One PO can contain several items, destinations and due dates. Use the PO line and schedule identifier, receiving site and material specification as the observation key, with each receipt linked beneath it. A header-level date cannot explain which part of a split delivery became usable.

Record timestamps with their timezone or declared site calendar. A midnight boundary can change an integer day difference even when elapsed time changes very little. Preserve original timestamps and calculate durations consistently, rather than editing times to match a rounded supplier promise.

The ABC XYZ segmentation guide can help prioritise the pilot materials. High operational consequence or unstable supply may justify more detailed investigation, even where annual purchase value is modest.

Calculate elapsed time from comparable events

Consider five invented replenishment observations measured in calendar days under the same start and availability definitions. Arrival durations are 8, 9, 7, 10 and 6 days. Additional receipt-to-release durations are 1, 3, 1, 4 and 1 day. Total usable-stock lead times are therefore 9, 12, 8, 14 and 7 days.

Mean arrival duration is 40 ÷ 5 = 8 days. Mean receipt-to-release duration is 10 ÷ 5 = 2 days. Mean total is 50 ÷ 5 = 10 days. The decomposition makes the average internal delay visible without blaming it automatically on the supplier.

For total lead time, deviations from the mean are −1, 2, −2, 4 and −3. Squared deviations sum to 34. Sample variance is 34 ÷ (5 − 1) = 8.5 days squared; sample standard deviation is square root of 8.5 = approximately 2.915 days. Use the sample denominator consistently and retain the five observations.

Five observations are too few to establish a robust planning distribution. This calculation demonstrates the record and arithmetic, not a recommended sample size. Show the observation count, date range and inclusion rules beside every reported mean or percentile.

Do not add standard deviations

For the same dataset, arrival-duration sample standard deviation is approximately 1.581 days and receipt-to-release sample standard deviation is approximately 1.414 days. Their sum, about 2.995 days, does not equal the directly measured total of about 2.915 days.

Component variances combine with their covariance, not by simply adding standard deviations. The sample covariance here is 2 days squared, so total variance is 2.5 + 2 + 2 × 2 = 8.5 days squared. When component delays rise together, ignoring that relationship can also understate uncertainty.

Measure the total directly when compatible timestamps are available, then use the components to investigate causes. Do not subtract an average inspection delay from every order and call the remainder a measured supplier duration. Actual order-level milestones preserve information that averages remove.

Give partial receipts an explicit completion rule

Suppose a 1,000-piece PO line receives 600 usable pieces on day 8 and 400 on day 14. Time to first usable receipt is 8 days; time to complete availability is 14 days. A quantity-weighted availability average is (600 × 8 + 400 × 14) ÷ 1,000 = 10.4 days.

Those three figures answer different questions. The weighted mean cannot support a claim that all 1,000 pieces were available on day 10.4. If a line launch needs 900 pieces at once, the first receipt does not meet that requirement. Retain the cumulative usable-quantity curve for such decisions.

Choose the measure appropriate to the replenishment model and display its name. For supplier scorecards, also keep agreed due quantities and milestones. Do not turn one split order into two equally weighted independent orders without considering how that changes the population.

Keep incomplete and exceptional orders visible

Open orders are not completed lead-time observations. Excluding them silently can bias a recent average downward because the slowest orders have not finished. Report completed-order statistics alongside an ageing view of open orders and explain the cutoff.

Cancelled lines, rejected receipts, replacement deliveries and emergency expedites need declared treatment. An expedited order may be useful for a separate recovery analysis while being inappropriate evidence for normal replenishment lead time. Keep the record and reason rather than deleting inconvenient observations.

Avoid pooling routes or material types with different processes merely to increase the count. Compare a stable supplier-material-site group over a stated period. If a process changes, identify the effective date and review whether earlier observations remain relevant.

Check the event sequence

Flag impossible sequences before reporting durations. A release time before arrival, a receipt quantity above the approved schedule without an explanation, or a later revision attached to an earlier commitment all require review. Preserve the source value and record the correction rather than forcing a negative interval to zero. A zero can be a legitimate same-time event, so it should not double as the silent replacement for missing or invalid timestamps.

Turn the record into a planning decision

The inventory policy guide supplies the policy context. Use validated availability lead time when reviewing the protection period, while keeping supplier execution measures separate for accountability. A buffer should cover a defined uncertainty, not an unexplained date-field mismatch.

Microsoft's reordering-policy documentation also highlights the role of calendars and expected existing supply. Verify how the local system treats nonworking days and receipts that are present but not usable. A corrected average in a spreadsheet has no effect if the planning engine still reads another parameter.

Start with one supplier and one repeat-use item. Reconstruct the last completed receipts and the current open schedule, resolve missing timestamps, and agree the partial-receipt rule. Only then compare alternative lead-time or safety-stock parameters using the approved replenishment model.

References

Oracle Replenishment Planning How Safety Stock Is Calculated release 26C

Microsoft Learn Business Central Handling reordering policies