Blog · Procurement and supply chain · Procurement
Supplier KPIs connect delivery, quality, lead-time reliability, and purchase cost to clear denominators. Work through one supplier's records and see which team owns each measure and its exceptions.
Supplier KPIs measure delivery, quality, purchase cost, and reliability using agreed business rules and the rows behind them. Start with a small balanced set rather than a long list of ratios. Each measure needs a denominator, a reporting period, an owner, and a decision it can change.
A key performance indicator earns its place when someone acts on it. For a supplier, the practical questions are whether goods arrive as promised, whether they are usable, whether lead time is predictable, and whether comparable items cost more.
Keep this framework distinct from a weighted supplier score. A delivery failure should remain visible even if the supplier has favorable prices. Likewise, dashboard layout is a separate task: the supplier performance dashboard shows how to present the measures once their definitions are settled.
| Measure | Definition and denominator | Export evidence | Business owner |
|---|---|---|---|
| OTIF | Complete on-time schedules / schedules due | Original promises and recorded receipts | Procurement with receiving |
| Rejection rate | Rejected units / inspected units | Inspection records and dispositions | Quality |
| Mean lead time | Completed lead days / completed observations | Release and completed receipt dates | Planning |
| Lead-time variability | Standard deviation of comparable completed lead times | Same completed receipt cohort | Planning |
| Matched-item price change | Current cost less baseline cost, divided by baseline cost, using current quantities | Invoice lines and approved baseline prices | Procurement with finance |
The procurement team KPI guide covers the wider purchasing function. The supply chain KPI guide extends into inventory and fulfillment. This framework stays at the supplier and site level.
Write a short definition for each KPI. For delivery, state whether the test uses purchase schedules, order lines, or full orders. Preserve the original due date and explain how buyer-approved changes are treated. Include overdue open schedules in the due-period denominator.
For quality, decide whether the rate measures rejected units, lots, or inspections. Those are different measures. If only selected units are inspected, disclose that coverage; an inspected-unit rejection rate does not automatically describe every unit received.
For lead time, choose calendar or business days and fixed start and end events. Keep uncompleted orders visible separately. A completed-order average can improve simply because the slowest open orders have not yet entered it.
For cost, match the item, specification, unit of measure, currency, and commercial scope. A price change on the same item differs from a change in average price caused by buying more expensive items.
A US manufacturer reviews Clearview Industrial at its Ohio site. The following schedules all cover the same item and unit, and each ordered quantity was eventually received and inspected. The received-by-due column records physical receipts against the original promise, before later inspection disposition. Delivery completion and quality rejection are separate tests here. All dollar amounts are USD, excluding freight and tax consistently in both price columns.
| Schedule | Ordered units | Received by due | Inspected units | Rejected units | Lead days | Baseline unit price (USD) | Actual unit price (USD) |
|---|---|---|---|---|---|---|---|
| A | 100 | 100 | 100 | 1 | 8 | 10.00 | 10.00 |
| B | 100 | 100 | 100 | 1 | 10 | 10.00 | 10.50 |
| C | 100 | 80 | 100 | 2 | 12 | 10.00 | 11.00 |
| D | 100 | 100 | 100 | 0 | 10 | 10.00 | 10.00 |
| Total | 400 | 380 | 400 | 4 | 40 |
These are illustrative business records, not an industry benchmark. Rates are calculated from unrounded inputs; rounding is stated with each result.
OTIF = Schedules complete by the promised date / Schedules due
Rejection rate = Rejected units / Inspected units
Schedules A, B, and D passed OTIF. Schedule C failed because only 80 of its 100 units were received by the due date. OTIF is therefore 3 / 4 = 75.0%.
The received-by-due quantity share is 380 / 400 = 95.0%. That quantity measure answers a different question; it cannot replace schedule OTIF. Show both only when the distinction supports a real decision.
Quality is 4 / 400 = 1.0% rejected units. Do not label this a lot failure rate: the underlying counted objects are units. Receiving owns the delivery evidence, while quality owns whether the recorded rejects were confirmed, reversed, or still awaiting disposition.
Mean lead time = Sum of completed lead days / Completed observations
Sample standard deviation = Square root of [Sum of squared differences from the mean / (Observation count - 1)]
Mean lead time is 40 / 4 = 10.0 days. Differences from that mean are -2, 0, 2, and 0 days. Their squared differences sum to 4 + 0 + 4 + 0 = 8, so sample standard deviation is the square root of 8 / 3, or 1.63 days rounded to two decimals.
NIST's measures of scale reference explains this sample standard deviation and why its units match the original observations. State that this is the sample formula; do not silently alternate between sample and population calculations.
The coefficient of variation expresses that spread relative to the mean:
Lead-time coefficient of variation = Sample standard deviation / Mean lead time
Using the unrounded standard deviation, the result is 16.3%, rounded to one decimal. NIST defines the coefficient of variation as this ratio and explains its measurement-scale limitations. Report the mean, spread, and observation count together. This short set is useful evidence for discussion, but too small to establish a stable expectation by itself.
Use actual quantities for both the actual and baseline extension. That holds volume and item mix constant for this comparison.
| Schedule | Actual extension (USD) | Baseline extension (USD) | Price difference (USD) |
|---|---|---|---|
| A | 100 × 10.00 = 1,000 | 100 × 10.00 = 1,000 | 0 |
| B | 100 × 10.50 = 1,050 | 100 × 10.00 = 1,000 | 50 |
| C | 100 × 11.00 = 1,100 | 100 × 10.00 = 1,000 | 100 |
| D | 100 × 10.00 = 1,000 | 100 × 10.00 = 1,000 | 0 |
| Total | 4,150 | 4,000 | 150 |
The matched-item price increase is (4,150 - 4,000) / 4,000 = 3.75%. The dollar check is $4,000 + $150 = $4,150. Reconcile the actual extension to the included invoice lines before calling this a supplier price movement. It is not automatically a realized savings figure or a complete landed-cost comparison.
The example shows incomplete delivery against the promise and a higher matched-item purchase price. Procurement can ask for the cause of the delivery miss and evidence supporting the price changes. Quality can review the rejected units and whether their disposition affected production. Planning can investigate lead-time consistency without confusing it with the average.
Compare the same measures by supplier, site, and period before escalating. If the concern is deterioration from an established baseline, use the supplier deterioration worked example rather than inventing a universal passing score.
Bring a sample of your own purchase schedules, receipt, quality, and invoice exports; Covirage maps the columns, and its tools compute and check the comparisons. The external AI model explains the findings with every figure traced to its rows. Create a free Covirage account, which takes about a minute, and set up your business with your own data.
See Covirage for procurement teams for the purchasing context.
Start with on-time-in-full delivery, rejected units divided by inspected units, lead-time average and variability, and matched-item purchase price change. Add measures only when a buyer can name the decision they support and the export that proves them.
OTIF tests whether each promised schedule was complete by its due date. A quantity fill rate measures the share of units supplied. A nearly complete late order may contribute many units to fill rate while failing the OTIF test.
Use common definitions where possible, but set targets against the contract, material criticality, and site requirements. A shared formula makes suppliers comparable; it does not establish that the same service threshold is appropriate for every purchasing situation.
Procurement should coordinate the definitions with receiving, quality, planning, and finance. Each measure needs an accountable business owner and a source owner who can resolve exceptions. Keep the approved definition and change history alongside the reported results.