Sign in

Blog · Territory, capacity and quota planning · Supply chain

Lead time variability per supplier per site: the number behind the safety stock

How a supply chain team measures actual lead time and its variability per supplier per site per item from purchase orders and receipts, why the spread drives safety stock more than the mean, the suppliers whose variability rose while their average held, and the inventory that a tighter supplier or a corrected lead time parameter would release.

The short answerActual lead time is receipt date minus order date per purchase order line, and its variability is the spread, the standard deviation or interquartile range, per supplier per site per item. Safety stock formulas use both, and the spread usually matters more. Compared with the planned lead time held in the ERP, the report shows the items where the parameter is wrong in either direction and the suppliers whose spread rose while their average held, and it values the inventory a corrected parameter or a tighter supplier would release.

A supply chain team sets safety stock from a lead time parameter in the ERP that was entered when the item was created. The receipts say what the lead time actually is, per supplier, per site, and how much it varies. This guide sets out actual lead time and its spread, the comparison to the parameter, and the inventory the difference is costing.

The measures

Per supplier, per site, per item, per period:

Actual lead time = receipt date − order date, per PO line Mean, median, interquartile spread, standard deviation Parameter gap = planned lead time in the ERP − median actual

Per item, per site:

Safety stock at current spread, at target spread, and the difference in units and value

The rows you need

  • Purchase order lines: PO, line, supplier, site, item, order date, quantity.
  • Receipts: PO, line, receipt date, quantity.
  • Item master: item, site, planned lead time, unit cost, service level.

Supplier identifiers only.

The assertion

every receipt joins to exactly one PO line, and receipt date ≥ order date

Receipts before orders, which are usually keying errors, fail it and are excluded from the lead time with a count.

A worked view

One item, two sites, same supplier.

Site Lines Median actual Spread (IQR) Planned Parameter gap Safety stock now At target spread Released
Plant A 48 21 days 4 days 21 0 380 units 380 none
Plant B 52 24 days 15 days 21 −3 1,410 units 420 990 units, $61,000

Same supplier, same item. Plant B's lead time is a little longer and four times as variable, and it is carrying a thousand units of safety stock that a steady supplier or a different lane would release. The planned parameter is also three days short, which is why Plant B expedites.

The supplier trend

Supplier Site Median now Median trailing 4 Spread now Spread trailing 4 Reading
S-0217 Plant B 24 23 15 6 Spread rose; average held
S-0442 Plant C 31 27 5 5 Average rose; steady

Supplier S-0217's average looks fine and its variability has more than doubled. That is the supplier conversation, and the on-time scorecard would not have raised it for another quarter.

Where it goes wrong

Supplier-level only. Plant B's problem is averaged into Plant A's.

Mean without spread. The safety stock driver is unmeasured.

Parameter never compared. The ERP plans on a number nobody has checked since the item was set up.

Released inventory reported as saving. It is an estimate at a stated service level. Say so.

Every month, per supplier per site per item

Mapped once, PO lines, receipts and the item master produce actual lead time, spread, the parameter gap and the released-inventory estimate every month. Covirage builds this from the exports as they are. The supply chain page describes the setup, and the OTIF trend guide covers the companion measure from the same receipts.

Questions people ask

Why per site?

The same supplier ships to different sites on different lanes with different carriers and receiving practices. A supplier's lead time to one plant can be steady and to another erratic, and a supplier-level figure averages the erratic one away.

How is the released inventory valued?

Safety stock at the current spread against safety stock at the target spread, for the same service level, times unit cost. The formula and the service level are stated on the report. It is an estimate of what a change would free, not a promise.

What about the planned lead time in the ERP?

It is a parameter someone set at item creation. The report shows planned against actual per item and site, and the items where planned is well below actual are the stock-outs waiting to happen; where planned is well above, the inventory is sitting for no reason.