Key takeaways
- A scorecard needs four dimensions: cost, quality, delivery and service, each with a written definition.
- Weights should reflect the part and the market, not a fixed template applied to every supplier.
- A score is only useful if it is tied to an action, a review date and a named owner.
- Measure the same way every period, and publish the method before the first score is issued.
What a scorecard is for
A supplier scorecard is a short, repeatable record of how a supplier performed against the things the buyer actually needs. It is not an audit, and it is not a substitute for a first article or a quality report. It is the periodic view that tells a programme whether the relationship is improving, stable or drifting.
The most common failure is a scorecard built to look thorough rather than to be used. It carries twenty metrics, most of which nobody can define, and it is issued once a quarter without a conversation. The supplier treats it as a form, and the buyer treats it as a filing requirement.
A useful scorecard is deliberately narrow. Four dimensions, a small number of metrics in each, a clear weight and a review meeting where the score is discussed with the supplier. The value comes from the conversation and the action, not the spreadsheet.
The four dimensions
Cost covers the price the buyer pays and how it behaves over time. It should include the piece price against the agreed basis, the accuracy of the quote against the final invoice, and the supplier's response to a should-cost or open-book review. It should not reward a low price that was achieved by cutting a specified process.
Quality covers conformity and the cost of non-conformity. Delivery covers the ability to meet the committed date and quantity. Service covers the less tangible behaviours: response time, documentation, engineering support, willingness to handle a change and honesty when something goes wrong.
The four are connected. A supplier with an excellent price and poor quality is not cheap. A supplier with perfect delivery and no engineering support can still slow a programme down. The scorecard should show the connection rather than hide it behind an average.
Metrics that can be defined
Each metric needs a definition, a data source and a period. If two people can calculate the same metric differently, it will be disputed at the first review. Write the definition down and keep it with the scorecard.
| Dimension | Metric | Definition | Illustrative weight |
|---|---|---|---|
| Cost | Piece price variance | Invoiced price against the agreed price for the period | 25% |
| Cost | Quote accuracy | Final invoice against the quoted price for new parts | 5% |
| Quality | Lots accepted first time | Lots accepted at goods in without sorting or concession | 20% |
| Quality | Non-conformance response | Days to contain and root cause a reported issue | 10% |
| Delivery | On time in full | Lines received on the committed date and in the full quantity | 25% |
| Service | Response and documentation | Working day response to a query and completeness of the document pack | 15% |
The weights are illustrative. They should be set for the part and the market, and agreed with the supplier before the first review.
Setting the weights
A generic weighting of 25% for each dimension is easy to explain and rarely correct. A part on a safety-critical assembly should weight quality more heavily. A commodity part in a competitive market should weight cost more. A part on a short lead time should weight delivery more.
The weights also signal behaviour. If cost carries half of the score, the supplier will optimise for cost. If quality carries half, the supplier will invest in inspection and process control. Neither is wrong, but the buyer should choose deliberately rather than inherit a template.
Keep the weights stable for at least a full review cycle. Changing them after the score is calculated destroys trust and makes the trend meaningless. If a weight must change, apply it from the next period and tell the supplier why.
Where the data comes from
Goods in supplies the quality and delivery data. Accounts payable supplies the price data. Engineering and procurement supply the service observations. The scorecard should draw on records that already exist rather than create a parallel reporting burden.
The weakest area is usually service, because it is a judgement. Turn it into observable events: days to respond to a query, number of document revisions, number of escalations, whether a corrective action was closed on the agreed date. A score built on events is easier to defend than a score built on impressions.
Where data is missing, mark it as missing rather than scoring it as zero. A zero punishes the supplier for the buyer's lack of records, and it distorts the trend.
Review cadence
Most suppliers need a monthly data pack and a quarterly conversation. The monthly pack keeps the numbers current and catches a drift before it becomes a dispute. The quarterly review gives enough time to discuss a trend and agree an action.
For a critical supplier or a new programme, monthly reviews are appropriate for the first few months. For a stable, low spend supplier, a six monthly review may be enough. The cadence should match the risk, not the calendar.
Send the pack before the meeting, not during it. The supplier should have time to check the data and prepare a response. A review where the score is revealed for the first time becomes an argument about the numbers instead of a discussion about the fix.
Treat the first period as a baseline rather than a verdict. Use it to agree the definitions and the data sources, and expect to correct a few records. A supplier who understands the method is more likely to accept a low score and act on it than one who sees the score as a surprise.
- 01
Publish the method
Share the metrics, definitions and weights before the first period starts.
- 02
Collect the data
Pull from goods in, accounts and engineering records, and mark any gaps.
- 03
Send the pack early
Give the supplier the score and the underlying data before the review.
- 04
Agree actions
Record what changes, who owns it and when it will be reviewed.
- 05
Follow up
Open the next review with the status of the previous actions.
Tie the score to a decision
A scorecard that never changes anything is a reporting exercise. The score should connect to a decision: a new enquiry, a volume shift, a development project, an audit frequency or a place on the next tender list.
Be explicit about the thresholds. A supplier below a defined score for two consecutive periods should expect a formal improvement plan. A supplier above a defined score should be considered first for new work. The thresholds do not need to be punitive, but they must be known in advance.
The same logic applies upward. If the buyer's forecasts, drawings or approvals cause a miss, that belongs in the review too. A scorecard that only measures the supplier will be treated as one sided, and a supplier is more likely to act on a score where it can see both sides.
Pitfalls to avoid
Do not average away a serious failure. A supplier with a perfect price and a field failure should not score well overall. Use a gate: a critical quality event caps the total score regardless of the other dimensions.
Do not compare scores across different part families. A high volume stamping and a low volume machined component have different risk profiles, and their scores are not on the same scale. Compare a supplier with its own history first, then with its peers in the same family.
Do not let the scorecard replace the relationship. The numbers start the conversation. The conversation is where the cause and the fix are found.
Frequently asked questions
What are the four supplier scorecard dimensions?
Cost, quality, delivery and service. Each needs written metric definitions, a data source and a period so that two people calculating the same score reach the same result.
How should the weights be set?
Set them for the part and the market. Weight quality more for safety-critical parts, cost more for commodities and delivery more for short lead time parts. Keep them stable for a full review cycle.
How often should a scorecard be reviewed?
A monthly data pack with a quarterly conversation suits most suppliers. Critical or new suppliers may need monthly reviews, while stable low spend suppliers can be reviewed six monthly.
What if the supplier disputes the score?
Show the underlying data and the written definition. If the data is wrong, correct it. If the definition is ambiguous, fix it for the next period rather than renegotiating the current score.
Sources and further reading
- ISO 9001:2015, Quality management systems, requirements
- Chartered Institute of Procurement and Supply, supplier performance guidance
- Qumek supplier review templates and programme notes
Qumek Trade Desk
Trade and logistics team
The team that handles Incoterms, customs classification, freight and landed cost for Qumek programmes.
Set up a scorecard for a live supplier
Send the part family and the metrics you already collect. The trade desk will reply with a draft metric set, definitions and a review cadence.
Request a draft scorecard