Plenty of manufacturers produce supplier scorecards. Fewer can point to a supplier whose performance improved because of one. The gap between measuring performance and improving it is where most scorecard programmes live, and closing it has more to do with how the scorecard is used than what’s on it.
Measure What You’d Act On
Scorecards bloat. Twenty metrics arrive, most are ignored, and the supplier can’t tell what matters. Four or five is enough.
On-time delivery, measured against the original agreed date rather than a revised one, and against your required date rather than their promise. This distinction changes the number substantially and it’s the one suppliers most often dispute — resolve the definition before you publish anything.
In-full delivery, tracked separately from on-time. Partial deliveries that arrive on schedule are still a problem, and combining the two hides which is happening.
Quality, as defect rate at receipt plus issues found later in production. The second part matters because problems discovered at assembly cost far more than ones caught at goods-in.
Responsiveness — time to quote, time to acknowledge, time to respond to a problem. Soft but predictive: it deteriorates before delivery performance does.
Price competitiveness against market movement, so you can distinguish a supplier passing on genuine cost increases from one widening their margin.
Weight By What it Costs You
An unweighted average treats a late delivery on a line-stopping component the same as one on a consumable with a month of buffer. Weight the metrics by the actual consequence for that category, and the score starts reflecting reality.
Why Scorecards Don’t Change Behaviour
Three common reasons, and they’re all fixable.
The supplier never sees it. A scorecard used only internally for sourcing decisions can’t drive improvement, because nobody at the supplier knows what to improve.
There’s no conversation. A monthly email with a number attached gets filed. A quarterly review where you walk through the data, ask what’s driving it, and agree specific actions with dates produces change.
There are no consequences in either direction. If good performance doesn’t lead to more volume and poor performance doesn’t lead to less, the score is information without incentive.
Look at Your Own Contribution
This is the part most programmes omit and it changes the tone of the entire relationship. Late purchase orders, frequent change requests, rush demands, unclear specifications, and slow payment all degrade supplier performance, and suppliers rarely raise these unprompted.
Ask. A supplier who’s late because you consistently order inside their lead time isn’t the problem you thought you had, and finding that out is worth more than another metric.
Keep it Light
Automated data, five metrics, quarterly reviews for your critical suppliers only. A programme that takes a week a month to run will be abandoned within a year.
Start with the ten suppliers where poor performance hurts most. Prove the review conversation produces change, then extend if it’s worth it. Most manufacturers find the top ten account for the overwhelming majority of the available improvement anyway.
ticktick.ai generates scorecards from transaction data with consequence-based weighting, including your own ordering behaviour against supplier lead times.
