AI Supply Chain BriefSupplier scorecard template: KPIs, weights and a worked quarterly example
Short answerA supplier performance scorecard tracks a handful of measurable KPIs, usually on-time delivery, quality (defective parts per million), corrective-action response and commercial accuracy, scored 1 to 5 against fixed bands and weighted to 100. Review it monthly or quarterly. Each total maps to an action: preferred, improvement plan or corrective action. Define every formula before the first score.
Fictional exampleExample Controls Co., its suppliers and every number in this example are made up for illustration. The score bands and weights are a sample policy, not a standard: set your own. Don't paste real supplier data, personal data or export-controlled data into an AI tool; see the data-safety checklist.
This page is about the ongoing performance scorecard: measuring suppliers you already buy from, every month or quarter. If you're choosing between bidders, that's selection scoring; use the weighted supplier scorecard prompt or Step 3 of the RFQ bid comparison example instead.
Which KPIs belong on a supplier scorecard?
Keep it to 4 to 6 KPIs you can pull from your own system every period without arguing about the data. These four cover most direct-material suppliers.
On-time delivery (OTD)
OTD % = PO lines received on time ÷ PO lines due in the period × 100. The formula is the easy part. The argument is always about what "on time" means, so write the window down before the first score: how many days early count as on time, how many days late, and whether a short shipment counts.
Published supplier requirements show how much this varies:
- DRiV's SupplyNet scorecard counts a delivery as on time from "one day early and zero days late to the due date and +/- 5% of order quantity".
- NORDAM's supplier rating quick reference allows up to 5 business days before or 1 business day after the PO line due date.
Same KPI name, very different windows. A supplier at 95% under one rule could be well below that under the other. These are examples of published requirements, not recommendations.
Quality PPM
PPM = parts rejected ÷ parts received × 1,000,000. Petoskey Plastics' supplier performance guidelines use this rejected-over-received formula. Decide whether you count rejects at receiving only or also later at the line, and whether a returned lot counts every part in it.
Corrective action (SCAR) response
Track the average days to close a supplier corrective action request (SCAR) and, if you can, how many were accepted on the first submission. Decide in advance what a supplier with no SCARs scores; in the example below it's a 5.
Commercial accuracy
Invoice price accuracy % = invoice lines matching the PO price ÷ invoice lines × 100. Price mismatches cost your accounts-payable team time even when they're small.
Optional extras
PPAP or first-article first-pass rate, audit findings, or responsiveness. Add one only if you can measure it the same way every period.
How do you score and weight it?
Turn each KPI into a 1 to 5 score with fixed bands, then weight the scores so the weights add up to 100. Here is the sample policy the example uses:
| KPI | 5 | 4 | 3 | 2 | 1 |
|---|---|---|---|---|---|
| OTD % | ≥ 98 | 95–97.9 | 90–94.9 | 85–89.9 | < 85 |
| Quality PPM | ≤ 50 | 51–250 | 251–500 | 501–1,000 | > 1,000 |
| SCAR avg days to close (no SCARs = 5) | ≤ 15 | 16–30 | 31–45 | 46–60 | > 60 |
| Invoice price accuracy % | ≥ 99 | 97–98.9 | 95–96.9 | 90–94.9 | < 90 |
Weights: OTD 35 · PPM 35 · SCAR 15 · Invoice 15 (total 100). Tiers: 4.0 or higher is Preferred, 3.0 to 3.99 needs an improvement plan, and below 3.0 means corrective action.
Real scorecards weight differently. DRiV splits Delivery 50% and Quality 50%, Petoskey Plastics uses OTD 40%, Quality 40%, Pricing Accuracy 10% and Continuous Improvement 10%, and NORDAM uses Delivery 45%, Quality 45% and Business Factors 10%. Pick weights that match what hurts your plant most, and keep them fixed for the year.
Worked example: a Q3 scorecard for three suppliers
Example Controls Co. (fictional) scores three suppliers of machined parts for Q3.
| Supplier 01 | Supplier 02 | Supplier 03 | |
|---|---|---|---|
| PO lines due / on time | 264 / 254 → 96.2% | 150 / 132 → 88.0% | 215 / 213 → 99.1% |
| Parts received / rejected | 25,000 / 3 → 120 PPM | 50,000 / 2 → 40 PPM | 20,000 / 29 → 1,450 PPM |
| SCAR avg days to close | 22 | No SCARs | 64 |
| Invoice lines / matched | 500 / 497 → 99.4% | 200 / 195 → 97.5% | 240 / 230 → 95.8% |
| Scores (OTD / PPM / SCAR / Invoice) | 4 / 4 / 4 / 5 | 2 / 5 / 5 / 4 | 5 / 1 / 1 / 3 |
| Weighted total (out of 5) | 0.35×4 + 0.35×4 + 0.15×4 + 0.15×5 = 4.15 | 0.35×2 + 0.35×5 + 0.15×5 + 0.15×4 = 3.80 | 0.35×5 + 0.35×1 + 0.15×1 + 0.15×3 = 2.70 |
| Tier | Preferred | Improvement plan (delivery) | Corrective action (quality) |
What the example teaches:
- Supplier 03 has the best delivery and the worst total. Parts that arrive on time but fail inspection still stop the line, which is why PPM carries the same weight as OTD.
- Supplier 02's parts are excellent but late. Its improvement plan should target delivery only. Talk to your planner too: late suppliers drive up safety stock (see ChatGPT prompts for inventory management).
- The total hides the story. Always show the KPI rows next to the total when you share it.
What do you do with each result?
- Preferred (4.0 and up): share the scorecard, say thank you, and keep them first in line for new work your policy allows.
- Improvement plan (3.0 to 3.99): agree on the one or two KPIs to fix, the actions and the dates, and review progress next period.
- Corrective action (below 3.0): issue a SCAR or corrective-action plan and start a sourcing review in parallel, so you have options if it doesn't improve.
How often? Many teams pull the data monthly and hold the formal review quarterly. User Solutions suggests a rolling 90-day window for operational monitoring and a 12-month trailing view for formal scorecard reviews.
For the quarterly review meeting itself, the Teams prompt in Copilot prompts for procurement turns the meeting recap into an action log. To keep an eye on supplier risk between reviews, see the supplier risk monitoring use case. For defect detection at receiving, see visual quality inspection.
Do it with AI
Export the period's receipts, rejects, SCARs and invoice matches, replace supplier names with codes, and paste them into your approved AI assistant with this prompt:
You are a supplier quality analyst. From the receipt and SCAR export below, calculate per supplier: OTD % (on-time lines ÷ lines due, on-time window [X days early / Y days late]), PPM (rejected ÷ received × 1,000,000), average SCAR days to close, and invoice accuracy %. Score each KPI with these bands [PASTE], apply weights [PASTE], and assign tiers [PASTE]. Show every calculation. Then draft a 120-word note to each supplier: what they did well, the one KPI to fix, and the date of the next review. Don't invent data; list any supplier with missing fields. [PASTE ANONYMIZED EXPORT: supplier code, lines due, lines on time, parts received, parts rejected, SCAR count, SCAR days to close, invoice lines, invoice lines matched]
What to check: recompute one supplier by hand, confirm the OTD window in the output matches your written policy, and confirm the weights add up to 100. Before you send a note, make sure it doesn't promise volume or threaten removal beyond your authority.
Free supplier scorecard template
Download the supplier scorecard template (CSV). It opens in Excel or Google Sheets and is pre-filled with the fictional Example Controls Co. data above. The OTD %, PPM, invoice %, scores, weighted total and tier columns are formulas, so you only type the counts.
Columns: supplier code, period, lines due, lines on time, OTD %, parts received, parts rejected, PPM, SCAR count, SCAR avg days, invoice lines, invoice lines matched, invoice %, a score per KPI, a weight per KPI, weighted total, tier and notes. Change the band thresholds in the score formulas to match your own policy.
Related: RFQ bid comparison example (for choosing suppliers) · ChatGPT prompts for procurement · How to respond to a supplier price increase
FAQ
What metrics go on a supplier scorecard?
Most direct-material scorecards use on-time delivery, quality measured in defective parts per million (PPM), corrective-action (SCAR) response time and invoice or price accuracy. Some add first-article or PPAP pass rate, audit findings or responsiveness. Keep it to 4 to 6 KPIs you can measure the same way every period.
How do you calculate on-time delivery?
Divide the PO lines received on time by the PO lines due in the period and multiply by 100. Define on time first: how many days early or late still count, and whether a short shipment counts. Published supplier scorecards use different windows, so write yours down.
What is supplier PPM, and what's a good PPM?
Supplier PPM is parts rejected divided by parts received, times 1,000,000. There is no universal good number: it depends on the part, the process and the industry, and your quality agreement with the supplier should set the target.
Is there a free supplier scorecard template in Excel?
Yes. The CSV on this page opens in Excel or Google Sheets, is pre-filled with the fictional example, and calculates OTD, PPM, scores, the weighted total and the tier for you.
What's the difference between a supplier scorecard and a supplier evaluation?
A supplier evaluation helps you choose a supplier, usually during an RFQ. A supplier scorecard measures a supplier you already use, period after period, on delivery, quality and commercial performance.
How often should you review a supplier scorecard?
Many teams collect the data monthly and hold a formal review quarterly. One option is a rolling 90-day window for day-to-day monitoring and a 12-month trailing view for formal reviews. Pick a rhythm your suppliers know in advance.
