SteelUnannounced audit readiness (batch 2 guide 3)
Factories rarely lie outright about capacity; the usual problem is optimistic math built on untested assumptions about shifts, yields, and subcontracting. This article shows how to decompose a capacity figure into checkable parts, verify each part against independent records, and agree on an audit scope that covers the weak points. The result is a capacity opinion you can actually schedule against.
Key takeaways
- Decompose every capacity claim into headcount, lines, hours, and yield before believing the headline monthly figure, a caution supported by acceptance sampling theory, which ties conclusions to defined plans rather than single specimens[1].
- Reconcile attendance, payroll, and production records against each other, because a single self-reported number proves nothing, a context shaped by technical barriers to trade rules, which expect technical requirements to stay transparent and non-discriminatory[2].
- Walk the subcontracted processes in person or include them explicitly in scope, since hidden outsourcing is the classic capacity trap, a caution supported by acceptance sampling theory, which ties conclusions to defined plans rather than single specimens[1].
- Count setup, changeover, and rework time in the capacity math instead of accepting nameplate output as available output, a context shaped by technical barriers to trade rules, which expect technical requirements to stay transparent and non-discriminatory[2].
- Check overtime patterns for sustainability rather than celebrating them, because peak output borrowed from rest days is not real capacity, a caution supported by acceptance sampling theory, which ties conclusions to defined plans rather than single specimens[1].
Many buyers commission capacity audits only after a late delivery, which means the audit starts as an argument about the past instead of a plan for the future. Shifting the timing earlier changes everything: the same fieldwork that assigns blame after the fact prevents the delay beforehand. This article is written for that earlier moment.
Early audits work because they surface constraints while options still exist: splitting the order, adjusting the ship date, qualifying a second line, or simplifying the product. Late audits merely document why none of those things happened. The cost of the audit is identical; only the timing decides whether it buys leverage or paperwork.
What follows is a practical sequence from scope negotiation through record cross-checks to corrective closure, with attention to the human side of audits. Tired suppliers, audit fatigue, and interview dynamics all affect what the auditor can learn. Accounting for those factors is part of professional audit management.
Which records actually prove capacity
Self-reported output totals prove little on their own, while mutually constraining records prove a great deal: attendance sets the labor envelope, material intake bounds the possible output, and shipment logs show what really left. Technical barriers to trade rules add the market-access angle: technical requirements should remain transparent and non-discriminatory across supplying countries.[2] Auditors triangulate these sources because fabricating all of them consistently is far harder than inflating one spreadsheet. Buyers should ask explicitly for triangulation rather than accepting a single total.
Payroll and time-clock data deserve emphasis since they are generated for purposes beyond impressing buyers, which makes them harder to stage. View details for the current service scope before you confirm the booking. Cross-checking those against production tickets for the same weeks closes most of the room for optimistic reporting.
Maintenance logs and downtime sheets complete the picture by explaining the gap between theoretical and demonstrated output. Acceptance sampling literature makes the same point from the other side: without a defined sampling plan, one result cannot carry the whole lot.[1] Factories that track stoppages honestly are usually managing them, while factories with suspiciously blank logs are usually hiding them. Either way the buyer learns something valuable about operational maturity.
Where capacity numbers usually mislead
| Audit area | What auditors review | Common gap | Buyer response |
|---|---|---|---|
| Bottleneck operation | Throughput of the slowest necessary step | Bottleneck never measured separately | Require a bottleneck-specific assessment |
| Subcontracted steps | External processes and their owners | Undisclosed outsourcing of key steps | Name each step and visit it |
| Yield and rework | First-pass yield and repair loop size | Nameplate output used as net output | Apply demonstrated yield to the math |
| Order mix effects | Changeover burden from product variety | Single-product rates applied to mixed orders | Ask for mix-adjusted estimates |
| Seasonal staffing | Hiring and attrition across the year | Peak-season crew assumed year-round | Get seasonal staffing in writing |
Use the matrix as a scope checklist: any row without corroborating records becomes explicit audit scope rather than background trust.
From scope letter to verified conclusion
- Walk the full flowFollow the material path including night shifts, rework rooms, and subcontracted steps that the paperwork minimized. Each discovered step gets its own record requests and observation time.
- Name the bottleneckMeasure the slowest necessary operation separately and cap the plan at what it can deliver. System output obeys the constraint regardless of how impressive the other stations look.
- Agree corrective actionsConvert findings into dated commitments with named owners at the closing meeting while attention is focused. Vague improvement intentions age into the next audit repeat findings.
- Verify under loadSchedule the follow-up during normal busy production rather than a staged quiet week. Fixes that hold under load are assets; fixes visible only under observation are liabilities.
- State the planning questionDefine the quantity, mix, and ship date the capacity must support before scoping any audit work. Audits scoped to abstract capacity produce abstract conclusions; audits scoped to a real order produce usable answers.
Recurring gaps in capacity evidence
Audit fatigue is real on both sides of the table, and tired people produce thin evidence and thinner conclusions. Acceptance sampling literature makes the same point from the other side: without a defined sampling plan, one result cannot carry the whole lot.[1] Buyers can reduce fatigue by consolidating overlapping audits, sharing recent reports across customers where appropriate, and scoping each visit to genuine open questions. Fewer, better audits beat a calendar crowded with ritual visits.
Corrective closure is where fatigue does its costliest damage, as follow-up visits get postponed and promises age into ambiguity. Technical barriers to trade rules add the market-access angle: technical requirements should remain transparent and non-discriminatory across supplying countries.[2] Fixed closure dates agreed at the closing meeting, with named owners on both sides, resist the drift that kills improvement. Each overdue week should trigger a written update rather than silent extension.
Verification visits earn their cost by checking that fixes survived contact with peak production rather than with audit-week conditions. Acceptance sampling literature makes the same point from the other side: without a defined sampling plan, one result cannot carry the whole lot.[1] Scheduling the verification during a normal busy period, not a staged quiet week, tests the corrective action honestly. Fixes that hold under load are assets; fixes that hold only under observation are liabilities.
A reminder before you commit the order
Before trusting any capacity figure, ask for the equation behind it and one month of corroborating records. Factories with real capacity show their math gladly; factories with aspirational capacity change the subject.
Sources
- Wikipedia: Acceptance sampling —— Overview of acceptance sampling including AQL concepts used in product inspection.
- WTO: Agreement on Technical Barriers to Trade —— WTO portal for the TBT Agreement covering technical regulations, standards and conformity assessment.
Frequently asked questions
How should subcontracted steps be audited?
Name each external step in the scope letter, visit the subcontractor or review their recent delivery performance, and confirm inspection rights and rejection records mirror the main factory. Undisclosed outsourcing follows someone else's schedule during crunch weeks. Backup options for critical external steps belong in the contingency plan.
What is the most common capacity illusion?
Nameplate output presented as available output, ignoring changeovers, rework, maintenance, and mix effects. Demonstrated output from comparable past months is the honest baseline. The gap between the two is where delivery promises go to fail.
How can buyers handle seasonal workforce swings?
Get the seasonal staffing pattern in writing with hiring and attrition data, then schedule critical production inside reliably staffed windows. Factories with stable crews deserve more scheduling trust. Avoid assuming the peak-season crew exists year-round without evidence.
When should overtime figures raise concern rather than confidence?
When high overtime is structural rather than seasonal, since it means quoted output already borrows from rest time. Trend the overtime across months instead of admiring a single compliant one. A factory living on overtime has no reserve left for rush orders.
// NEXT MOVE
Stop eating pages at 2 AM. Let SteelRats absorb them.
Senior platform engineers embedded in your stack — Kubernetes, Terraform, observability, incident response. Fixed monthly retainer. No juniors. No surprises.