The verification, audit, tooling and switching playbooks all end at the same sentence: keep better records. This page is the keep-better-records machine. Six KPIs — quality, punctuality, paperwork, price stability, communication, reserve depth — computed quarterly from data you already hold, weighted into a composite, mapped to four action bands (stay / fix / double-track / exit). Two consecutive quarters in a band makes the trigger real. The point is not math: it is that a scored panel tells you which relationship to invest in before the failure quarter happens — and the option-to-switch worth more than every switch, per the switching playbook.
Why Feel Doesn't Scale Past Three Suppliers
Below three suppliers, memory works. Above it, feelings cluster late: the factory that shipped a bad Q4 remembers it shipped great Q1-Q3, remembers no blame at all, and your year-end "gut says something's off" has no receipts. Meanwhile the pattern that actually predicts season-killers — slippage in weeks 8-11 of the cycle, paperwork first-pass decay, quote-creep on spec-fixed items — is plainly visible in records you already hold:
- PSI reports (from the inspection protocol) carry defect rates per batch — quality data exists the moment you run your first PSI.
- PI trail + shipping docs carry promised vs actual ship dates — punctuality data exists on every order.
- Your inbox carries reply timestamps and first-pass document rates — communication data exists silently.
The scorecard's only demand: tabulate four times a year instead of reconstructing once a crisis. Two hours per quarter, per supplier panel.
The Six KPIs
| KPI | Source (data you already hold) | Score rule (0-100) |
|---|---|---|
| Quality | PSI major-defect rate vs your AQL budget | 100 = half budget; 80 = at budget; 60 = 1-2× budget; floor 20 = rejected batch |
| Punctuality | Ship-date slippage in working days, actual vs PI | 100 = on-time; -4 pts/day; floor 20 = >10d or silent reschedule |
| Paperwork | First-pass rate of docs (PI, packing list, cert copies, HS codes) without your correction | 100 = all first-pass; 40 = you correcting every shipment |
| Price stability | Quote drift per quarter on UNCHANGED specs | 100 = ±1%; -10 pts per further 1% drift not tied to BOM/material index |
| Communication | 24h reply rate on the named working channel, queries needing decision | 100 = ≥95%; 40 = ghost weeks on hard questions only |
| Reserve depth | Verified backup capacity of the panel itself (this KPI grades YOU, not them) | 100 = qualified switch-ready alternate per the sprint; 0 = single-source |
The sixth KPI is the game-changer and deliberately inverted: it scores your own side. A panel with no qualified alternate caps every other relationship's leverage — the credible option beats the executed switch, but only if the option exists.
Weights and the Composite Score
Default weights for custom-manufacturing relationships: quality 30, punctuality 20, paperwork 15, price stability 15, communication 10, reserve depth 10. For repeat-stock replenishment, shift to quality 25 / punctuality 30 — calendar slots dominate in stock goods. Weights change quarterly at most; tossing weight mid-quarter to justify a feeling corrupts the record's value.
| Example (Q3, custom product) | Raw | ×weight |
|---|---|---|
| Quality — 1.4% major vs 1.25% budget | 76 | 22.8 |
| Punctuality — 3d average slippage | 88 | 17.6 |
| Paperwork — 90% first-pass | 90 | 13.5 |
| Price stability — quote +3% on unchanged BOM | 70 | 10.5 |
| Communication — 96% in 24h | 100 | 10.0 |
| Reserve depth — alternate mid-sprint | 60 | 6.0 |
| Composite | 80.4 → Band B (fix-tier) |
Read the worked example like a manager: the composite says fix-tier, the line-breakdown says where. Price stability is the wiggle line — +3% on unchanged specs is quote-creep, answerable with the triangulation set, not yet a structural complaint.
Score Bands and the Actions They Trigger
| Band | Trigger (2 consecutive quarters) | Assigned action |
|---|---|---|
| A — 85+ stay-and-grow | Composite ≥86 twice | Ask for the volume-price step; share forecast to anchor capacity; this is where annual unit upgrades get granted cheerfully. |
| B — 70-84 fix-tier | Composite in band twice | Named weakness with dated corrective items — the CAP discipline from the audit playbook, applied relationship-side. One quarter of no movement escalates to C. |
| C — 55-69 double-track | Composite in band twice | Open the qualified alternate: run the 21-day sprint, move 20-40% volume per the dual-track table. Not punishment — insurance with a deadline. |
| D — below 55 exit path | Composite in band twice (or one quarter with hard flags: payment games, subcontracting discovery, entity instability) | Start the switching playbook's exit discipline immediately: clean invoices, transfer tooling per clauses, keep the bridge, never exit holding their tooling or cash hostage. |
Hard flags shortcut the two-quarter rule — payment-account changes (see payment methods), concealed subcontracting, addressed this page nowhere but the switching gates: they are not scores, they are exits wearing scores. Do not average your way past them.
The Panel Structure: A / Bench / Watch
Scorecards rank; panels structure. The mature panel runs three tiers by product line:
- A-bench (1-2 suppliers): 60-80% of volume. Gets your forecasts, your volume commitments and your quarterly review meeting. A-bench relationships earn spec improvements for free — the factory that knows your annual plan quotes tighter.
- Useful bench (1-2): 20-40% volume, sprint-qualified within the last 2 quarters. The dual-track destination. If your bench tier is empty, every scorecard line above caps out at leverage-parity — fund it before year-end.
- Watch list (0-1): candidates under verification or ex-A-tier on probation. Sees sample orders only. Their scorecard runs too, from whatever data exists.
The structure question to ask annually: has volume drifted without merit? Volume creep to a mid-band supplier because "they always say yes" is the most common panel failure mode among buyers we debrief — and the scorecard quarter that catches it pays for itself.
The Half-Year Review Meeting
Twice a year, 45 minutes per supplier, with the composite and line-breakdown printed. The flow that works:
- Share the composite and the top positive. Chinese suppliers respond to numbers shared early and calmly — the meet-the-numbers-first culture documented in the field guide. Lead with what's working.
- Name the weakest line, ask what it looks like from their side. Quote-creep lines surface BOM-reality causes half the time (copper index, packaging substrate) — and become genuinely negotiable items.
- Set the next quarter's one lever. One number to move, their chosen path. A supplier who commits to "punctuality 3d → 1d with pre-approval now written into our PI process" has co-owned the scorecard; the one who deflects has too.
- Document both versions. Their narrative and your record, one page. Disagreement documented now is a solved problem later; disagreement discovered in crisis is a lawyer problem.
Review meetings are also where reserve-depth projects announce themselves: "we're qualifying a second source for continuity" said at review time reads as maturity, not threat — said only at exit time reads as betrayal (the switching playbook's smaller-than-it-looks rule again).
Frequently Asked Questions
Which six KPIs go on a supplier scorecard?
Quality (PSI major-defect rate vs AQL budget), punctuality (ship-date slippage in days), paperwork (document first-pass rate), price stability (quote drift on unchanged specs), communication (24h reply rate on the working channel), reserve depth (qualified switch-ready capacity on your side). All six are computed from data you already collect; the scorecard just tabs them quarterly.
What score triggers action on the scorecard?
Bands, not single scores: 85+ stay-and-grow, 70-84 fix-tier with dated CAP items, 55-69 open a qualified second source, below 55 start the exit path. Two consecutive quarters in band arms the trigger — except hard flags (payment games, subcontracting discovery), which shortcut straight to exit.
How much work does quarterly scoring take?
Two hours per quarter per panel once tabulated: PSI defect rates are on inspection reports, ship dates on the PI trail, reply rates in your inbox. The only new discipline is doing it on schedule instead of rebuilding after a crisis. First quarter takes half a day to set the spreadsheet; after that it's a fill-in.
Should I share scores with my suppliers?
The composite and line-breakdown, at the half-year review: yes. Raw internal judgments, single-quarter dips, and the existence of other bidders: no. Sharing works when it is a standing routine (numbers shared early read as management maturity); sharing first-time in a dispute reads as an attack catalogue.
What weights should quality vs punctuality get?
Custom manufacturing: quality 30 / punctuality 20. Repeat-stock replenishment: punctuality 30 / quality 25 — calendar slots dominate stock goods. The weights rarely matter more than the bands as long as quality+reliability together exceed 50% and price stability stays at or under 15 — over-weighting price is how panels drift to mediocre-cheap suppliers.
Does the scorecard replace an annual factory audit?
No — the 2-3 year re-audit clock from the audit ladder still applies for custom-manufacturing partners. The scorecard tells you WHICH relationships earned the next audit bill and where the audit's corrective-energy should aim. Score, audit, visit — the three layers read production reality together (the visit plan covers the culture facts scores can't).
Want your panel scored by a second pair of eyes?
Tell us your product lines and current supplier roster. We'll reply within 24-48h with where leverage sits, which tier is missing, and what the next quarter's one lever should be.
Request Panel Review →