Supplier scorecards: measuring delivery, quality and claim resolution fairly

Image: AI-generated illustration of a supplier-performance review; the documents are illustrative examples.

The supplier with the lowest price is not always the one your team wants to order from. Perhaps its deliveries need recounting, its product codes keep changing or every credit takes weeks of chasing. A scorecard can make that work visible. But if nobody can explain how the score was calculated, it becomes one more report to argue about. Start with the problems people actually recognise.

Select the problems that affect your operation: late arrivals, incomplete orders, wrong products, defects or slow claim resolution. Avoid adding a measure simply because software can display it. Each measure needs a clear question, such as whether an order arrived by its confirmed date or whether the accepted quantity matched the order.

Keep price separate from operational performance at first. A single combined score can hide the reason a supplier is difficult to use. Review price, delivery, quality and support alongside one another before deciding how to weight them. The priorities for an urgent replenishment line may differ from those for a long-planned seasonal purchase.

Choose what you are counting before choosing how to display it. Nine correct lines and one late line can produce very different results when measured by order, line or shipment. Each view may be useful, but switching between them will make a comparison misleading. Explain the basis in ordinary language and apply it consistently.

For a simple order-level measure, define on-time and complete using the agreed date and the agreed quantity. Record an approved partial-delivery arrangement separately. Preserve the original commitment and any agreed revision so the score is not improved by moving the deadline after the delivery has already failed.

Use order acknowledgements, receipt records and claim references rather than recollections. A low score should be traceable to actual transactions. If a supplier disputes it, your team should be able to identify the affected orders and review whether the definition was applied consistently.

Suppose six of eight orders arrived on time and complete. That is 75% on an order-level measure, but “six of eight” is worth showing beside the percentage. The sample is small, and the two failures may share one fixable cause. A score should open that conversation, not pretend to settle the whole relationship with a number.

How well did they put it right?

Measure how the supplier resolves the problem as well as how often it occurs. An acknowledgement is not a credit, and a promise of replacement is not stock back on your shelf. Record when a properly documented claim began and when the agreed resolution was completed. Keep the still-open cases visible.

Separate supplier-caused problems from delays caused by your own missing information. If your team sends an incomplete claim, record that fact. A fair scorecard encourages both sides to improve; it should not attribute every problem in the buying process to the supplier.

Discuss repeated causes with the supplier using examples. Wrong pack quantities may need clearer data, while missed dates may require more realistic capacity commitments. Agree an owner, an action and a review date. A promise to communicate better is weaker than a defined change to when order confirmation is sent.

Decide what your business will do while improvements are tested. That might mean smaller commitments, extra inspection or a second source for critical lines. Match the response to the consequence of failure. Avoid applying the same rule to every supplier regardless of product importance or substitutability.

Keep the scorecard stable enough to compare periods, but check whether it still measures the problems that matter. A new fulfilment arrangement or product range may introduce a different service requirement. Document any changed definition and avoid comparing a new measure directly with an older one as if nothing changed.

End the review with a change somebody can carry out. Which confirmation will arrive earlier? What packing detail will be corrected? Who will check whether it worked? A few well-defined measures backed by real orders can support a much better buying decision than a colourful league table. The point is to make the next delivery better, and to know what your business will do if it is not.