Sampling levels used in watch inspection: part 2

Key takeaways

1. Calibrate the sampling level per order to the cost of an escaped defect instead of reusing one default for every shipment.
2. Formal acceptance testing logic applies at the factory gate: verify finished goods against agreed criteria before transfer ^[1].
3. Draw destructive-test pieces as a recorded sub-sample inside the main sample and report them as separate lines.
4. Documented quality management practice expects sampling decisions and acceptance rules to be written and traceable ^[2].
5. Review the plan after every inspection so levels track supplier history rather than habit.

The companion guide to this article explained what sampling levels are; this one explains how to use them strategically across a real watch program. Most buyers set one level once and forget it, applying the same scope to launch orders and mature repeats, to flagship automatics and commodity straps. That habit wastes inspection effort where risk is low and starves it where risk is high.

A better approach treats the sampling plan as a living decision reviewed per order. New suppliers earn stricter scrutiny until their history justifies relaxation; new movements and complications draw tighter levels until field data accumulates; accessories riding alongside flagship pieces can carry lighter levels without distorting the verdict. Special levels for slow or destructive tests flex independently of the main visual sample.

This guide shows how to calibrate levels to risk, combine general and special levels in one inspection, and write plans that survive supplier disputes. It is written for sourcing managers and brand owners running repeat watch programs rather than single trial orders.

From default level to deliberate choice

The default Level II serves average risk well, which is exactly why it misleads buyers into treating it as universal. A launch order for a new automatic from an untested factory carries multiples of the risk of a twelfth repeat of a proven quartz model, yet both sail through the same sample size when nobody revisits the plan. Deliberate choice starts by scoring each order on three axes: supplier history, product complexity, and commercial consequence of failure. High on any axis argues for Level III; low on all three may justify Level I for accessories or even for the main goods.

Cost arithmetic supports this flexibility more often than buyers expect. Doubling the sample roughly doubles the inspection time on site, while a single escaped systematic fault can cost a full rework cycle plus air freight plus review damage. Spending the extra inspector hours on high-risk orders is inexpensive insurance; spending them on proven low-risk repeats is waste. View details and compare the scope against the risk profile of the specific order before confirming the booking.

Calibrating the level to supplier history

New suppliers start strict because no history means no basis for trust, however convincing the samples looked. The first two or three shipments deserve Level III on the main goods with tightened attention on the failure modes from the pre-production meeting. As consecutive inspections pass with findings confined to scattered minors, the buyer earns the right to step down to Level II with documented justification. This progression rewards good factories with lower inspection friction while keeping the evidence trail intact.

Deterioration must move the dial the other way without sentiment. Two consecutive inspections with major findings in the same family, such as repeated hand-alignment faults or recurring coating failures, justify tightening back to Level III until the factory demonstrates correction across full shipments. Documented quality management practice expects sampling decisions and acceptance rules to be written and traceable ^[2], so record each change with its reason. Suppliers respect buyers whose strictness visibly follows evidence rather than mood.

Combining general and special levels in one visit

1. Sample the lot firstDraw the general-level sample at random across all cartons and pallets first, since it represents the whole consignment. Run the visual, functional, and measurement checks on these pieces before anything is consumed by destructive testing.
2. Sub-sample for slow testsSelect immersion, tape-test, and endurance pieces by further random draw from within the general sample, using the agreed special level. Mark each consumed piece and record its position so counts stay exact.
3. Report separatelyKeep destructive findings on separate report lines with their own sample sizes, never folding two immersion results into hundred-piece visual arithmetic. A failed immersion pair condemns the sealing process on its own evidence.
4. Segregate tested goodsExclude consumed pieces from shippable quantities and confirm their disposition with the factory before cartons are resealed. Tested watches are normally scrapped or returned, never quietly reintroduced.
5. Read the plans togetherCompare findings across the two plans for a joined picture: scattered visual minors plus passing immersion suggests process noise, while clean visuals plus failing immersion points squarely at the sealing station. Direct correction accordingly.

Writing plans that survive disputes

A sampling plan earns its keep in the meeting after a failed inspection, when the factory argues the sample was unfair or the rules were unclear. Plans that survive name the standard edition behind the code letters, the general and special levels, the AQL per defect class, the definition of each class for this product, and the disposition of tested pieces. Formal acceptance testing logic applies at the factory gate: verify finished goods against agreed criteria before transfer ^[1], and watches deserve that logic in writing rather than in memory.

Review the plan after every inspection instead of filing the report unread. Track findings by family across shipments to spot the drift that justifies tightening, and track clean runs that justify relaxation. Buyers who maintain this loop find their inspection spending falling over time while their defect escape rate falls faster, because effort concentrates where evidence says it matters.

Frequently asked questions