How to Audit Manufacturing and QA Before PO Release
2026/05/23

How to Audit Manufacturing and QA Before PO Release

A buyer-side field guide for auditing CTQ control, measurement reliability, and lot traceability before committing volume.

A factory can ship excellent pilot samples and still fail during mass production.

When that happens, root cause is usually not machining capability itself, but weak control logic: unclear CTQ checkpoints, loose reaction plans, or traceability that looks good on paper only.

The First Question I Ask on Audit Day

"Show me how you detect drift before customer complaints."

If the answer depends on manual experience only and cannot be tied to a written control plan, the system is fragile.

One-Day Audit Plan That Works

TimeWhat to VerifyEvidence You Should See
09:00-10:30Process flow for your part familyRouter, station sequence, CTQ marks
10:30-12:00Measurement system reliabilityGauge list, calibration status, CMM method
13:30-15:00In-process controlInspection frequency + reaction rules
15:00-16:30Final release and traceabilityOutgoing checklist + lot demo
16:30-17:30Open risks and closure planOwner, due date, verification method

Remote audit can work, but only if you can see live records and station execution, not slides only.

CTQ Control Plan: Minimum Useful Format

CTQMeasurement MethodFrequencyReaction Plan
Output shaft runoutCMM + dial indicator cross-checkStart-up + interval checkStop lot, segregate WIP, fixture re-verify
Flange datum positionCMM datum methodPer batch + after tool changeQuarantine affected lots and root-cause within 24h
Backlash proxy featureFunctional jig + dimensional proxyDefined by risk tierBlock release until engineering review

A CTQ table without reaction plan is just documentation, not risk control.

Measurement Reliability: What Buyers Miss

Many teams ask, "Do you have CMM?" That is the wrong question.

The better question is, "How do you know this measurement system is trustworthy for this CTQ?"

Common industrial gating references:

  • Gauge R&R <= 10% for critical characteristics is usually treated as strong.
  • 10-30% is often conditional and requires written risk acceptance.
  • Beyond 30% usually means the measurement system needs redesign before relying on the data.

Use these as decision references, then adjust for your own field-risk tolerance.

Capability and Drift Rules Before PO

For critical fit features, many buyer teams set a pre-ramp expectation such as Cpk >= 1.33.

What matters even more than the exact number is whether drift triggers are explicit:

  • Who gets alerted when trends move toward spec limits?
  • What containment action starts immediately?
  • How fast must corrective action be closed?

If this is unclear, the PO risk is operational, not theoretical.

Traceability Stress Test (Do This Live)

Ask the team to pick one outgoing lot and demonstrate, in real time:

  1. Material lot linkage
  2. In-process inspection record
  3. Final release record
  4. Revision baseline used for that lot

If they cannot complete this chain quickly and consistently, future field incidents will be expensive to isolate.

Escalation SLA Template (Buyer Friendly)

Use a simple written SLA:

  • Acknowledge issue within 24h
  • Provide containment plan within 72h
  • Close corrective action within 30 days (or agreed project timeline)

The SLA is only meaningful if each severity level has a named owner.

PO Release Scorecard

AreaWeightRelease Signal
CTQ control maturity30Control plan + reaction logic complete
Measurement reliability20Critical measurements are trustworthy
Traceability depth20Lot-to-material-to-revision chain proven
Escalation readiness15Owner + SLA + closure logic defined
Change governance15Effective-lot and notice policy is clear

Practical rule: do not release PO below 75/100, and do not release at any score if traceability fails live demo.

If you need a pre-PO audit checklist tailored for humanoid joint projects, contact [email protected] or WhatsApp +86 18857971991.