How to scope quality control for mixed SKUs: part 2
A mixed-SKU order should never be inspected as one blended lot. Different products carry different defect risks, different test steps, and different acceptance limits, so one pooled sample cannot represent any of them fairly. The fix is to split the order into homogeneous inspection lots first, then sample and judge each group on its own terms. Buyers who scope this way get verdicts they can actually act on, instead of averages that hide the failing variant.
Key takeawaysSplit mixed-SKU orders into homogeneous inspection lots before sizing any sample, because sampling logic assumes uniformity within the lot [1]. · Group SKUs by shared materials, processes, defect risks, and acceptance criteria rather than by carton or container convenience. · Confirm any equipment-level or function-level test steps per group in advance, following factory acceptance test practice of checking against agreed requirements [2]. · Give every group its own accept or reject verdict so one failing variant cannot hide inside a blended pass. · Write the lot split, sample basis, and per-group criteria into the booking so the inspector and the factory work from the same map.
Mixed-SKU orders are where inspection plans most often quietly fail. A buyer consolidates ten product variants into one shipment, books a single inspection, and receives a single verdict. The report says the goods passed, yet one variant keeps generating customer complaints. Nothing in the report explains why, because the plan never distinguished the variants in the first place.
The root cause is statistical, not procedural. Sampling logic assumes the inspected lot is reasonably uniform, so a sample drawn across unlike products cannot represent any of them well. A defect rate that would fail a single-product lot can dissolve into an acceptable blended average when diluted across nine healthy SKUs and one troubled one.
This guide shows how to scope quality control for mixed SKUs by splitting the order into sensible inspection lots before anyone discusses sample sizes. It covers grouping logic, per-group sampling, the testing steps that differ by group, and a booking format that keeps each group's verdict separate and actionable.
What goes wrong with a single blended verdict
Picture an order with ten SKUs sharing a container. Nine are simple, stable products from a mature line. One is a new powered variant with a complex assembly. A single sample drawn proportionally across all ten will contain mostly units from the stable products, because they dominate the quantities. The new variant contributes a handful of units, too few for its defect rate to move the blended result.
The inspector follows the plan, the blended math passes, and the shipment sails. Weeks later the powered variant fails in customers hands while the report on file says everything was fine. The report was not dishonest; it answered the wrong question. It described the average of ten unlike things instead of the condition of each thing.
Scoping each product group as its own lot changes the question back to something answerable. UTS third-party product inspection can then apply the right sample basis, the right checks, and the right acceptance rule inside every group, and report a separate verdict the buyer can act on without punishing the healthy SKUs for one group's failure.
The cost objection rarely survives arithmetic. Per-group scoping adds inspection time, but far less than the rework, air freight, or chargebacks that follow an escaped defect in a blended lot. Buyers who have lived through one hidden-variant failure seldom return to blended sampling.
| Grouping approach | How lots are formed | When it fits |
|---|---|---|
| By product family | One lot per family sharing materials and process | Stable catalog with clear families |
| By risk level | New or complex SKUs split out from mature ones | Orders mixing proven and new items |
| By test requirements | Powered or regulated items grouped by test set | Orders with mixed testing needs |
| By production line | One lot per line or subcontractor | Split production across sites |
| By acceptance limits | Tight-tolerance items separated from general goods | Mixed critical and non-critical defects |
Sampling and testing logic inside each group
Sampling assumes uniformity, so create it first
Acceptance sampling is the practice of judging a lot from a representative sample under defined accept and reject rules. Its mathematics assumes the sampled units come from a reasonably uniform population, and its standard references describe the operating characteristics of such plans [1]. A mixed carton of unrelated products violates that assumption, which is why the lot split must come before any discussion of sample sizes.
Within a homogeneous group, the familiar machinery works as intended. The buyer sets the inspection level, reads the sample size for the group quantity, classifies defects into critical, major, and minor, and applies the acceptance numbers. Repeat this per group and the order yields several small, honest verdicts instead of one large, misleading one.
Confirming test steps before inspection day
Some groups need more than visual and dimensional checks. Powered functions, safety features, and custom tooling outputs each need defined test steps with pass criteria, performed on the right units, with the right fixtures. Factory acceptance testing follows the same principle at equipment scale, verifying against agreed requirements before handover [2].
The scoping document should therefore list, per group, which functional tests run, which instruments are required, and whether any samples go to a laboratory. Confirming this while scoping avoids the familiar inspection-day discovery that the needed fixture is at another site or that nobody agreed which standard the test follows.
Scoping a mixed order in six steps
- Inventory the SKUsList every SKU with quantities, materials, processes, and test needs. Flag new items, powered items, regulated items, and anything from a different line or subcontractor as candidates for separation.
- Form homogeneous lotsCluster SKUs that share defect risks and acceptance criteria into inspection lots. Keep together what fails together; separate what must be judged on different terms.
- Set per-lot rulesFor each lot, state the inspection level, the defect classification, and the acceptance numbers. Record which functional tests and measurements apply to that lot alone.
- Check readiness per lotConfirm the production and packing status per lot, because lots finish at different times. Schedule visits so each lot is inspected when it is genuinely ready, not when the container is.
- Book with the lot map attachedWrite the lot map, the per-lot criteria, and the separate-verdict requirement into the booking. Share it with the factory so carton marking and staging match the inspection plan.
- Decide per lot, not per containerRelease, rework, or hold each lot independently. A failing lot gets corrective action without blocking healthy lots, and repeat failures in one lot point to its specific process.
Review the lot map after every mixed order and adjust the groupings. SKUs that always pass together can merge; groups that keep diverging should stay split. The scoping document is a living tool, and each order should make the next one cheaper to plan.
Sources
- Wikipedia: Acceptance sampling —— Overview of acceptance sampling including AQL concepts used in product inspection.
- Wikipedia: Factory acceptance test —— Overview of factory acceptance testing performed before shipment or installation.
Frequently asked questions
Why does one blended sample fail for mixed-SKU orders?
When one sample covers several different products, a defect concentrated in a single SKU can hide inside an otherwise acceptable overall result. The blended verdict then releases a bad SKU alongside good ones. Scoping each product group as its own inspection lot keeps failures visible where they actually occurred and lets buyers accept, rework, or hold per group.
How should buyers group SKUs into inspection lots?
Group SKUs that share materials, processes, defect risks, and acceptance criteria. Splitting every color or size variant into its own lot usually wastes inspection time, while grouping unrelated products hides real differences. A practical rule is to separate groups where the defect types, the test steps, or the acceptable limits genuinely differ.
How does acceptance sampling apply inside each SKU group?
Acceptance sampling examines a representative portion of a lot and applies a defined accept or reject rule to the whole lot. Its logic assumes the lot is reasonably uniform, which is exactly why mixed orders must first be split into homogeneous groups. Once each group forms its own lot, standard sampling tables and defect classifications apply cleanly within it.
Where does factory acceptance testing fit into mixed-SKU scoping?
Factory acceptance testing checks equipment or systems against agreed requirements before delivery or use, often at the supplier premises. For mixed-SKU programs with custom tooling, fixtures, or powered functions, confirming those test steps per product group during scoping prevents inspection-day surprises about which tests apply to which items.
You've read the take. Now argue it.
1.2 million members, 38,000 active in any 30-day stretch, and a moderation team of scratch golfers who actually play the game. Register once, post forever.
Join the Conversation →