Verifying overtime records: what buyers miss

Key takeaways

1. Separate system findings from one-off lapses in the report, since only systemic gaps predict how the next order will run.
2. Close every corrective action with evidence and a deadline, and verify closure on the next visit rather than accepting paper promises.
3. Treat worker interviews as sampling, not interrogation: small, random, private conversations reveal whether procedures live on the floor or only in binders.
4. File the audit report with the sourcing decision it supports, so the next buyer or auditor inherits context instead of starting blind.
5. Use the audit to price risk into the contract — payment terms, QC intensity, and order allocation should all reflect what the visit actually found.

Overtime auditing connects labor standards to production reality: hours data constrains output claims, output data constrains hours claims, and pay data must satisfy both. The triangle either closes cleanly or exposes the fabrication.

Third-party evaluation matters because the auditor has no order to protect and no relationship to preserve. That independence is what lets uncomfortable findings reach the buyer intact, which is the entire point of paying for the visit.

The buyers who extract most value from audits run them as a program, not events: initial evaluation, corrective closure, periodic re-verification, with each round narrower and sharper than the last.

What compliance requires

Voluntariness and consent complete the picture: overtime must be voluntary with documented worker agreement, and pregnant, nursing, or juvenile protections observed where applicable. Coerced overtime documented as voluntary fails twice — hours and honesty.

Rest-day guarantees bound the whole structure: weekly rest entitlements, holiday arrangements, and annual leave records. Factories meeting daily limits while denying rest days comply with the letter's shadow and miss its substance.

Monthly and daily hour limits set the ceiling: auditors reconstruct actual hours from attendance traces and compare against statutory caps for the locality and period. Systematic exceedance across sampled months is a system finding, not a seasonal excuse. The sampling logic behind these checkpoints follows the acceptance-sampling principle: a defined sample can stand in for the lot when the plan, sample size, and acceptance number are fixed before inspection begins.^[1]

What happens on site

Document review runs in parallel with observation wherever staffing allows: licenses and certificates, quality manuals, recent production and QC records, corrective logs. Dates, signatures, and consistency across documents matter more than the polish of any single binder.

Interviews close the triangulation loop: managers describe the procedure, records show whether it was followed, and workers confirm whether it is real. Gaps between the three are findings, not misunderstandings to be smoothed over on the spot. Once the booking is confirmed, the coordination team at UTS books the visit once the factory confirms the goods are ready.

The closing meeting states findings plainly with severity levels before anyone leaves the site. No finding should first appear in the written report — surprise findings signal an auditor who avoided the hard conversation, and factories rightly distrust them. Under the WTO TBT framework, solid conformity evidence is what keeps differing national regulations from becoming shipment-stopping surprises.^[2]

Overtime evidence checks

All four legs need passing grades; strength in three never compensates fabrication in the fourth.

Reading the result commercially

Share the corrective section with the factory promptly and require a dated response plan. Audit value decays fast: findings delivered weeks later meet changed personnel and faded memory, and closure rates collapse.

Feed the findings into contract terms for approved factories: tighter incoming checks where the audit found weakness, staged volume where systems are young, explicit re-audit triggers where history is short. The report should change the commercial terms, not just the file.

Sort findings by consequence, not by count. One systemic gap — incoming material never verified, overtime records unreliable — outweighs a dozen minor housekeeping notes, and the sourcing decision should weight them accordingly. Laboratory credibility rests on accreditation, and the ILAC mutual recognition arrangement is what lets an accredited test result travel across borders.^[3]

Language discipline matters during interviews. Questions route through the auditor's interpreter, never through factory management, and answers are recorded privately. Relayed interviews through interested parties are not evidence.

Evening factory-dormitory observations, where the scope includes them, reveal working-hour realities that daytime records may disguise. Entry and exit patterns at shift change are hard to stage and easy to read.

Subcontracting discovered mid-audit should be mapped immediately: which processes, where performed, under whose control. Undisclosed outsourcing found by accident usually indicates more undisclosed outsourcing not yet found.

The auditor's independence needs visible protection on site: no gifts beyond courtesy, no factory-arranged transport that creates obligation, findings discussed with the buyer first. Small courtesies accepted casually become leverage later.

Senior factory attendance at the opening meeting predicts cooperation quality all day. When the owner or general manager attends, records appear faster and awkward areas open sooner; when only a junior handler appears, expect a managed tour.

Buyer habit that pays off

Frequently asked questions