Supplier performance reporting

A supplier scorecard should fit on one page. Delivery reliability, quality, lead time, price, and responsiveness cover most of what operations needs. Compliance, such as certifications and audit results, works better as a gate than as a weighted score.

More measures do not produce a better supplier. They produce a longer meeting and weaker agreement about what matters.

Start with the measures that change a decision. If a low score would not lead to a conversation, an order change, or a development plan, the measure is reporting for its own sake.

On time means the receipt happened within the agreed window, not that the supplier dispatched on time. In full means the full ordered quantity arrived. Measure at line level, then aggregate to order level only if partial deliveries are unacceptable in that category.

Write the rules down: the window in days, whether weekends count, how early deliveries are treated, and how substitutions are scored. A supplier arguing about definitions during a review is a sign the definitions were never agreed.

A practical window is the agreed delivery date plus or minus the tolerance written in the contract. Early delivery can be as disruptive as late delivery when it arrives before the warehouse has space or before payment is due. Record who agreed the date: the supplier confirmation, the purchase order, or a later change, because most disputes about lateness are really disputes about which date was promised.

Quality needs its own measure. Defect rate = rejected units / received units. Add the cost of poor quality where it is material: rework, sorting, expediting, and credit claims. A cheap supplier with a high defect rate is not cheap.

Price variance = (invoice price - purchase order price) / purchase order price. Track it alongside delivery because a supplier that holds price but misses dates still disrupts production.

Quality data often sits outside the purchasing system. Inspections, customer complaints, and production rejections may live in separate logs, so agree which source is authoritative before publishing a score.

Assign weights before the first review and keep them for at least four quarters. Example: OTIF 40, quality 25, lead time 15, price 10, responsiveness 10. Score each measure from 0 to 100, multiply by the weight, and add.

A supplier scoring 88 on OTIF, 92 on quality, 70 on lead time, 85 on price, and 80 on responsiveness would score 85.2. Use bands: A above 85, B from 70 to 85, C below 70. The arithmetic is simple; the discipline is in not changing the weights after the result.

Use a rolling twelve-month window for the headline score and a three-month view for early warning. A single bad month should be visible without dominating the annual rating. If a measure has no data for a period, score it as not applicable and reweight the others rather than entering zero, which would punish the supplier for a reporting gap.

A score without a consequence is decoration.

Thresholds should be realistic for the category. A custom fabricator and a commodity distributor cannot meet the same delivery standard, so set the target by category and review it when the supply market changes.

Consequences should escalate in steps. A first miss gets a conversation, a repeated miss gets a plan, and a persistent miss changes the sourcing decision. Skipping steps makes the process feel arbitrary.

Share the data pack a week before the meeting. The supplier should see the same numbers, the same definitions, and the same exceptions. Surprises turn a review into an argument.

End with an action log: what changes, who owns it, and when it will be checked. Review the previous log first, every time. Suppliers learn quickly whether the process is real.

Keep the scorecard visible between reviews. A quarterly document that disappears for three months cannot influence day-to-day expediting decisions.