How to Measure Supplier Performance Without Damaging the Relationship

Here is what happens without measurement. The supplier who let you down badly once is remembered for years. The supplier who has been a day late on a third of deliveries all year is not remembered at all, because no single delivery was dramatic.

Measurement fixes that. It turns supplier conversations from memory and mood into facts both sides can see. Done well, suppliers usually welcome it, because it also shows when they are doing a good job.

Done badly, it produces a spreadsheet nobody reads, arguments about whose numbers are right, and suppliers who learn to hit the target while missing the point. The process below is designed to avoid those outcomes.

#What "supplier performance" actually covers

Five things, for most companies:

Delivery

Did it arrive when agreed, in the quantity agreed? This is usually the first measure worth having and the easiest to get from your own records.

Quality

Was it right? Defects, returns, rework, services that had to be redone.

Commercial accuracy

Did they charge what was agreed? Invoice price against contract price is one of the most useful and least measured checks.

Responsiveness

When something went wrong, how fast did they respond and fix it?

Risk and relationship

Financial health, continuity plans, and whether they bring ideas or just fill orders.

APQC, the benchmarking organisation, singles out two supplier measures as standard: average supplier lead time and the percentage of deliveries on time. If you measure nothing else, start there.

#Step 1: Decide which suppliers to measure

Not all of them. Measuring 400 suppliers properly is impossible for a normal team, and measuring them badly is pointless.

Pick the suppliers that matter most. A sensible starting list is your largest suppliers by spend plus any supplier you depend on for something critical, even if spend is small. For most mid-sized companies that is ten to twenty-five suppliers.

The classic way to sort suppliers comes from Peter Kraljic's 1983 Harvard Business Review article: rate what you buy by its effect on profit and by how risky its supply is. Suppliers of high-impact or high-risk items get measured closely. Suppliers of routine, easily replaced items can be managed with a lighter touch.

#Step 2: Choose a few measures, not many

The UK Government's Sourcing Playbook gives unusually direct advice on this. Having too many performance measures — which it defines as more than 10 to 15 per service — "will lead to overcomplicated contracts and ambiguity with suppliers".

For most suppliers, three to five measures are enough. A starting set:

Notice the last column. Every one of these can come from records you already have, if your purchase orders and goods receipts are recorded properly. That is deliberate. Measures that rely on someone remembering to fill in a survey do not last.

#Step 3: Define each measure so nobody can argue

This is the step most companies skip, and it is the one that decides whether measurement works.

Take OTIF, the most common supplier measure. McKinsey studied it across the consumer goods industry with the Trading Partner Alliance and found there is no standard definition. Does on time mean the date you requested, or the date the supplier promised? Within a specific delivery slot, or any time that day? Is "in full" measured by whole order, by line, or by individual case?

Each choice gives a different score for exactly the same deliveries. Of the companies McKinsey surveyed, 92% agreed an industry standard would create value — which tells you how much confusion the lack of one causes.

Before measuring anything, write down:

• Which date counts — requested, confirmed, or the date on the purchase order.
• How much tolerance applies — same day, within one day, within the delivery window.
• Whether early counts as on time. It often should not, for reasons below.
• What "in full" means — every line, or a percentage of the order.
• How partial deliveries and substitutions are scored.
• What is excluded — for example, delays you caused by changing the order.

#Step 4: Agree the measures with the supplier

A scorecard the supplier has never seen is not a performance tool. It is a grievance file.

The UK Sourcing Playbook is explicit that misunderstanding how measures work "can result in unintended outcomes and / or service failures", and that it is important to work with suppliers "to ensure KPIs are jointly shaped and understood".

There is a practical reason too. Buyers and suppliers genuinely see the same events differently. In RapidRatings' 2026 risk survey, 66% of buyers reported supply disruption while only 35% of suppliers did. Agreeing definitions up front is how you avoid spending review meetings arguing about whose version of events is right.

#Step 5: Build a simple scorecard

A scorecard is just one page per supplier showing each measure, the target, the actual result and the trend. Keep it plain. An illustrative example:

Two tips. Show the trend, not just the latest number, because direction matters more than a single quarter. And resist combining everything into one overall score — a supplier averaging 90% might be perfect on quality and failing on delivery, and the average hides exactly what you need to discuss.

#Step 6: Review on a rhythm that matches importance

Send the scorecard before the meeting, not during it. The meeting should be about what to do next, not about reading numbers aloud.

#Step 7: Use the results to improve, not just to punish

Penalties have a place. Walmart is the best-known example. In 2017 it began fining suppliers for late and early deliveries, and from September 2020 it required 98% on-time, in-full delivery across all categories, with a fine of 3% of the cost of goods on shipments that missed.

But penalties change behaviour in ways you may not want. McKinsey found about 25% of consumer goods deliveries arrived more than two hours before their scheduled slot. Suppliers were avoiding late-delivery penalties by arriving early — which disrupts the receiving warehouse and leaves trucks standing idle. The measure improved. The operation did not.

That is why early delivery should usually count as a miss, and why penalties work best alongside improvement plans rather than instead of them. When a supplier misses target, the useful questions are: why, what will change, and by when?

The UK Sourcing Playbook puts it well: contracts should be designed to incentivise delivery of the things that matter, "to minimise perverse or unintended incentives and to promote good relationships".

#Step 8: Look beyond delivery once a year

For your most important suppliers, add an annual check on the things monthly data does not show: financial health, continuity plans, and whether the relationship is producing ideas as well as orders.

Olivier Berrouiguet, chief executive of Synertrade, described the goal on the Art of Procurement podcast: "You need to know your suppliers, what they can and what they can't do, and what they're ready to do on top of the relationship you have every day."

#If you are ISO 9001 certified, you already need this

Clause 8.4 of ISO 9001:2015 requires organisations to set criteria for evaluating, selecting, monitoring and re-evaluating the external providers they depend on, and to keep records of it. Many certified companies meet the letter of this with an annual approved supplier list review. A proper scorecard meets it far better, and gives you something useful at the same time.

#Does it make a difference?

The evidence says yes. In Deloitte's 2025 survey of more than 250 chief procurement officers, 84% of the leading organisations met or beat their plan for supplier performance. Among the rest it was 59%.

The same survey shows where leaders are putting effort. Increasing collaboration with suppliers was one of the three strategies CPOs expected to deliver most value, chosen by 38%. And 61% focused on more information sharing with suppliers as a way to manage risk. Measurement is what makes that collaboration concrete: it gives both sides the same facts to work from.

#Six common mistakes

1. Measuring every supplier, and therefore measuring none of them properly.
2. Tracking only price, and missing the late deliveries and invoice errors that cost more.
3. Relying on surveys and memory instead of purchase order and receipt data.
4. Never showing the supplier their own scorecard.
5. Using the numbers only to penalise, and teaching suppliers to game the measure.
6. Setting targets once and never revisiting them as the relationship matures.

#Where to start

Pick your ten most important suppliers. For each, choose three measures — on-time, in-full delivery, quality, and invoice accuracy is a good default. Write the definitions down. Share them with each supplier. Run the numbers for last quarter from your own records.

That first run will almost certainly show you something you did not know. It usually also shows how much depends on your own records being accurate — which is often the first thing to fix.