Supplier Scorecard Metrics That Drive Improvement
Short answer: A supplier scorecard drives improvement only when it scores four things — quality (PPM and lot acceptance), delivery (on-time-in-full), responsiveness (quote and issue-closure time), and cost behaviour (price stability and cost-down delivery). Weight them roughly 40/30/15/15, score monthly, and review quarterly with a written action plan for any line below target. Suppliers that receive the same scorecard every month for two quarters typically cut defect PPM by a third and late deliveries by half. If a supplier cannot tell you their own PPM and OTIF numbers, that is your first finding — and a reason to quote the part elsewhere.
Why most supplier scorecards fail to change anything
Most scorecards die in a spreadsheet. A buyer collects data, colours some cells red, emails a PDF, and nothing happens. The supplier either never opens the file or disputes the numbers because the definitions were never agreed.
The scorecard is not a report. It is a management system with three moving parts: agreed definitions, a fixed cadence, and consequences. Remove any one and the whole thing becomes decoration.
Three failure patterns show up again and again in sourcing teams:
- Too many metrics. A 30-line scorecard gets skimmed. Six to eight metrics, weighted, is the practical ceiling.
- Metrics the supplier cannot influence. Scoring a Chinese supplier on your own forecast accuracy teaches them nothing.
- No consequence ladder. If a red score changes nothing for three quarters, the supplier learns the scorecard is theatre.
A workable scorecard fits on one page, uses numbers both sides can pull from their own systems, and links every red score to a named action with a date.
The four core metric families
Quality: PPM is the headline, not the whole story
Parts per million defective is the number everyone quotes, but PPM alone hides a lot. A supplier shipping 200,000 parts a year with 400 PPM looks better than one shipping 2,000 parts with 400 PPM, even though the second has a far higher risk of a single bad lot reaching your line.
Pair PPM with two supporting numbers:
| Quality metric | Definition | Typical target | Why it matters |
|---|---|---|---|
| Defect PPM | Rejected units ÷ total received × 1,000,000 | < 500 PPM | Headline quality signal |
| Lot acceptance rate | Accepted lots ÷ total lots received | > 98% | Catches small-volume risk |
| First-pass yield (supplier-reported) | Good units ÷ units started | > 96% | Predicts future PPM |
| SCAR closure time | Days from issue to verified fix | < 14 days | Tests whether fixes are real |
| Repeat defect rate | Defects recurring within 90 days | < 5% | Exposes paper corrective actions |
The repeat defect rate is the one most teams skip, and it is the most revealing. A supplier that closes a SCAR in five days but repeats the same defect three months later is writing documents, not solving problems.
Delivery: OTIF, plus the reason code
On-time-in-full (OTIF) should be measured against the original confirmed date, not the revised one. If you allow date revisions to reset the clock, OTIF becomes meaningless.
Define "on time" precisely. A common and defensible rule: shipment arrives at your dock within the agreed window, with the full ordered quantity, correct revision, and complete documentation. Anything short of that is a miss.
Then track the reason code. Late because of raw material, late because of capacity, late because of a quality hold, and late because of a documentation error are four different problems with four different fixes. A supplier at 88% OTIF with all misses caused by one plating subcontractor is in far better shape than one at 88% with misses spread randomly across five causes.
Responsiveness: quote speed and issue closure
This is where sourcing teams in China often have a real advantage, and where you should measure it explicitly.
| Responsiveness metric | Definition | Typical target |
|---|---|---|
| Quote turnaround | Working hours from RFQ pack to priced quote | < 24 h (BQUQ: 12 h) |
| Engineering response | Working hours to answer a technical query | < 24 h |
| Sample lead time | Days from PO to first article | 7–15 days |
| Issue acknowledgement | Hours to acknowledge a complaint | < 8 h |
| Corrective action plan | Days to a written root-cause plan | < 7 days |
Quote turnaround is a leading indicator of everything else. A factory that cannot price a part quickly usually cannot plan capacity quickly either. When you compare suppliers, a quote comparison guide helps you normalise what each quote actually includes before you score it.
Cost: stability and cost-down, not just price
Price is the metric buyers love and the one that causes the most damage when used alone. Score cost behaviour instead:
- Price stability: how many unplanned price changes occurred in the period, and were they justified by documented material indices?
- Cost-down delivery: did the supplier hit the agreed annual productivity commitment?
- Quote accuracy: how far did the final invoice land from the quoted price?
- Cost-avoidance ideas: how many design-for-manufacture suggestions did the supplier raise unprompted?
That last one is underrated. A supplier who suggests changing a bend radius to eliminate a secondary operation is worth more than one who shaves 2% off a unit price.
Weighting the scorecard
Weights should reflect your business, not a template. A medical-adjacent or safety-critical part pushes quality higher; a fast-moving consumer product pushes delivery higher.
| Metric family | Weight (typical) | Weight (quality-critical) | Weight (speed-critical) |
|---|---|---|---|
| Quality | 40% | 55% | 30% |
| Delivery | 30% | 25% | 40% |
| Responsiveness | 15% | 10% | 20% |
| Cost behaviour | 15% | 10% | 10% |
Score each family 0–100, multiply by weight, and sum. Then apply a simple banding:
- 85–100 — Preferred: first look for new business, eligible for longer-term agreements.
- 70–84 — Approved: normal business, watch the lowest-scoring family.
- 55–69 — Conditional: new business paused, written improvement plan required within 30 days.
- Below 55 — Exit track: dual-source immediately, phase out over one to two quarters.
The numbers matter less than the consistency. A supplier who sees the same definitions and weights for four consecutive quarters will start managing to them.
Making the scorecard drive real improvement
Share the raw data, not just the score
A supplier who receives "78/100" learns nothing. A supplier who receives a line-item breakdown showing 1,240 PPM in March driven by burrs on a stamped part, plus the three lot numbers involved, can act the same week.
Share the underlying data — inspection records, non-conformance reports, delivery logs — in a consistent format. If you buy stamped parts, defects often trace back to tooling wear or strip feed issues, and your supplier needs the specific lot and cavity to investigate. The same logic applies to custom metal stamping and to machined parts where a single worn insert can affect a whole batch.
Set one improvement target at a time
Suppliers cannot fix four things at once, especially smaller factories running lean teams. Pick the single lowest-scoring family and agree one measurable target for the next quarter. "Reduce burr-related rejections from 1,240 PPM to under 600 PPM by 30 June" is a target. "Improve quality" is not.
Run a quarterly business review with a fixed agenda
A 45-minute call, same agenda every quarter:
1. Scorecard walk-through (10 min)
2. Supplier's own root-cause view on the two worst lines (10 min)
3. Open actions from last quarter — closed or not (10 min)
4. Next quarter's single improvement target (10 min)
5. Commercial topics: capacity, new projects, tooling (5 min)
Keep it to that. Reviews that drift into price negotiation stop being improvement meetings and start being avoided.
Audit the scorecard itself once a year
Ask three questions annually: Are we measuring anything the supplier cannot control? Are the weights still right for our product mix? Are we actually applying the consequence ladder, or have we been carrying a 58-score supplier for two years?
That last question is the uncomfortable one. Most scorecard failures are not data failures — they are discipline failures on the buyer's side.
Where scorecards interact with capacity and risk
Scorecards measure the past. They do not tell you whether a supplier can absorb your growth. Pair the scorecard with a capacity conversation — how many shifts, how many machines, what utilisation, what lead time at 2× volume. Our notes on supplier capacity assessment cover the questions worth asking.
Similarly, a supplier can score well and still carry structural risk: single-source raw material, one critical machine, one skilled operator. A scorecard is one input into a broader risk mitigation plan, not a replacement for it.
For first-time suppliers, run the scorecard alongside a structured vetting checklist for the first two quarters. Once the relationship stabilises, the scorecard alone carries the load.
A note on scoring source-direct factories
Factories that quote direct — rather than through a trading layer — tend to score differently. Responsiveness is usually better because the engineer quoting is the engineer who will run the job. Quality data is usually available but less polished; you may need to agree a reporting format rather than receive one.
At BQUQ, four production lines run under one roof in Dongguan — CNC machining to ±0.005 mm, metal stamping, custom springs, and heat sink production — under ISO9001. That matters for scorecards because a defect can be traced to the process that caused it without a chain of intermediaries. Quotes go out within 12 working hours, MOQs are flexible, and we will share PPM and OTIF data in whatever format your scorecard uses.
If you are building a scorecard and want a supplier who will report against it, send the RFQ pack to sc@bquq.com.
Frequently Asked Questions
Q: How many metrics should a supplier scorecard have?
A: Six to eight weighted metrics across four families is the practical ceiling. Beyond that, buyers skim and suppliers cannot prioritise. If you need more granularity, keep the headline scorecard at eight metrics and hold the detail in an appendix that only gets opened when a score goes red.
Q: Should I score suppliers monthly or quarterly?
A: Score monthly, review quarterly. Monthly scoring keeps the data current and catches drift early; quarterly reviews give suppliers enough time to actually fix something before the next conversation. Monthly review meetings burn goodwill and rarely produce change because there is no time for corrective action to show results.
Q: What is a good OTIF target for a Chinese supplier?
A: 95% OTIF against the original confirmed date is a reasonable target for established relationships, with 98% achievable for repeat, stable-volume parts. New suppliers or new tooling often sit at 85–90% for the first two quarters. Anything below 85% consistently usually points to a capacity or planning problem rather than a logistics one.
Q: How do I handle a supplier who disputes my quality data?
A: Agree the measurement method before you start scoring — gauge, sampling plan, acceptance criteria, and who arbitrates. If disputes persist, request a joint first-article inspection on the next lot and document the result. Most disputes come from different measurement methods, not from dishonesty, and they disappear once both sides use the same gauge.
Q: Can a scorecard be used with a single-source supplier?
A: Yes, and it is more important there. When you cannot switch suppliers quickly, the scorecard becomes your early-warning system. Weight responsiveness and cost-avoidance ideas higher, since you cannot use competitive pressure, and set improvement targets that build capability inside the existing relationship rather than threatening to move the work.
Related Resources
- About BQUQ and our four Dongguan production lines: /about/
- CNC machining to ±0.005 mm: /cnc-machining/
- Metal stamping and tooling: /custom-metal-stamping/
- Custom compression springs: /compression-springs/
- Sourcing and industry trends: /industry-dynamics/
- Full technical article library: /bquq-blog/
- Common sourcing questions: /faq/
- Contact the engineering team: /contact/
Authored by the BQUQ Engineering Team. BQUQ (Dongguan) runs CNC machining (±0.005 mm), metal stamping, custom springs, and heat sink production in one ISO9001 factory. Source-direct from Dongguan, China — quote in 12 hours: sc@bquq.com | WhatsApp +86 13713157787 | www.bquq.com
