Supply Chains
What a Supplier Audit Can and Cannot See
A factory audit is a sampled snapshot taken on an announced day, which explains why audits catch documentation problems reliably and hidden subcontracting far less often.
Buyers send auditors into supplier factories to verify that a plant is what the contract says it is. The method is sampling on a scheduled day, and that design decides what the audit can find.
An audit is a sample, not a census
An auditor spends a day or two in a plant that runs continuously all year. What is observed is a thin slice of production, chosen partly by the auditor and partly by the host.
Sampling works well for conditions that are stable. Fire exits, machine guarding, calibration records and material certificates look much the same on any day the plant is running.
Sampling works badly for conditions that vary. Overtime patterns, peak-season staffing and the handling of a rush order are precisely the things a single visit is unlikely to land on.
Announced visits change what is there to observe
Most audits are scheduled, because an auditor needs the right managers present and the line running. Notice is practical, and it also gives the site time to prepare.
Preparation is not necessarily deception. A plant that tidies its records and briefs its staff is behaving the way any organisation does before an inspection.
Unannounced visits remove that preparation window, which is why some buyers use them as a supplement. They cost more, waste trips when the line is idle, and strain the relationship.
Paperwork travels better than practice
Documents are the easiest thing to verify and the easiest thing to construct. Time records, wage ledgers and training logs can be complete and internally consistent without describing what happened.
Auditors cross-check paper against physical evidence: production output against hours claimed, canteen throughput against headcount, electricity use against shifts worked. Inconsistencies between independent traces are more informative than any single record.
This is why experienced auditors ask for records the site did not expect to produce. A prepared set of documents is coherent with itself but rarely with everything else.
The work that leaves the building
The hardest thing for an audit to detect is production the audited site never performs. Work sent to an unapproved subcontractor happens somewhere the auditor has no right of entry.
Signals exist. Output that exceeds what the installed machines could produce, finished goods with no corresponding work-in-progress, or gaps in the internal movement of materials all point outward.
Following those signals requires a second visit to a site the buyer did not know about, which is why unauthorised subcontracting is usually found through incidents rather than through scheduled audits.
What the result actually certifies
An audit report says that on a particular day, against a particular checklist, a sample of conditions met a defined standard. That is a narrower claim than the certificate implies.
Buyers who treat the report as a continuous guarantee are extending it beyond its evidence. Buyers who treat it as one input among order data, complaint records and site visits get more from it.
Standards and audit protocols also change over time, so two reports issued years apart may not be measuring the same things even when both look clean.
Questions readers ask
Is just-in-time discredited?
No, but its preconditions are better understood. It performs well with short reliable lead times and less well where supply variability is high, which is a statement about context rather than a verdict.
How much buffer is the right amount?
It depends on demand variability, supply variability and the service level chosen. There is no universal figure, and applying one target across a whole catalogue usually wastes money and misses sales at once.
