A kit can contain correctly made components and still differ from the agreed packing list. In an illustrative example, a pack specified to contain 20 clips contains 18, or the quantity is correct but the items belong to the wrong variant. Check identity and quantity separately; neither confirms the other.
Before accepting a checked pack, establish what the method shows about its contents and record any mismatch for rechecking.
Define the check unit before counting
A pack-count check only means something once the unit being checked is fixed. Two buyers can open the same carton and reach opposite conclusions because one is checking a retail pack and the other is checking the inner bag inside it.
Record these inputs before applying a checking method:
| Field | Actual entry or reference | Why it changes the check |
|---|---|---|
| Kit / variant / pack revision | Identifies the agreed configuration being checked. | |
| Exact unit being checked | Distinguishes one loose item, inner bag, retail pack or carton from another check unit. | |
| Component reference and verified marking or geometry | Defines how the required part is distinguished from another variant. | |
| Expected quantity and unit | Sets the agreed count and whether it means pieces, bags, packs or cartons. | |
| Applicable agreed colour / label / instruction revision | Identifies the reference for each included item; a count alone does not confirm the right version. | |
| Packing layer and component position | Shows where the component belongs within the specified unit. | |
| Method and source / validation record | Links the planned observation to its reference and records the limits of the conclusion. |
Complete the entries for the actual pack. This input record fixes what will be checked; the mismatch record below captures what was found and how it was rechecked.
A BOM can serve as that reference, but the point here is not to build one. It is to hold a specific, agreed version steady long enough to check against it.
What a difference shows, and what it does not
A count difference is evidence about the units that were opened. It does not, on its own, tell you how many units in the batch carry the same error, and it does not diagnose a cause. A shortage, a wrong variant and a wrong colour can reach the packer through different routes, and a single discrepancy number that lumps them together hides the others.
Two observations have to be kept apart. Quantity is how many pieces are present. Identity is whether those pieces are the right part, colour and revision. A pack with the right count can still be wrong, and a pack with a wrong count can still be the right part. A method that only counts is not checking identity, and a method that only weighs is not checking either one completely.
Reference to hold, and the boundary with sample approval
The agreed specification or approved reference unit is the comparison baseline. Which units and modules get opened, how many, and how a finding is handled is set by the buyer or the appointed inspection team against the process they work to. A checked unit does not establish the whole lot, so the reference has to be paired with the sampling decision rather than replacing it. This article does not prescribe sample sizes or quality levels; those belong to the inspection plan. Verify only the cards, bags and accessories that the pack specification actually contains. Not every kit includes a vine clip, a card or a bag.
What each checking method establishes
| Method | Useful result | Conditions needed | Blind spot | Next action when uncertain |
|---|---|---|---|---|
| Manual layout and direct count | Exposed items counted; visually distinguishable identities recorded | Items separated so each is visible; a confirmed component reference to compare against | Does not prove hidden material, performance, sealed inner content or lot conformity | Reconcile the ambiguous item against the reference, then recount |
| Weight-based screen or count estimate of one confirmed component type | Gross and net mass recorded against a defined reference piece mass | One homogeneous component type; tare or packing condition defined; reference piece mass and its variability known; scale resolution and repeatability adequate for the difference being judged | A mass match alone does not identify colour, wrong variant, card version or exact contents | Return to a direct count or another method agreed as verified |
| Sealed-module check | Modules counted; label, lot and position reconciled | Module identity, lot and pack position known; a scoped content-verification step in the plan | A label and an intact seal do not prove the identity or count of the pieces inside | Review the module evidence and verify selected contents under the agreed plan. |
| Combined checks | A record showing which observation supports identity, quantity or screening | Each contributing method's own conditions met | Combining records does not produce a zero-error guarantee | State which conclusion each observation supports before relying on it |
Manual layout is the method that puts a human eye on each piece, so it can catch a wrong variant that happens to be the same size and mass as the correct one. That value depends on the variants being visually distinguishable in the first place. Where two variants look alike or a component's material is hidden, appearance settles nothing, and time and error depend on the actual process rather than a fixed rate.
Relying on weight for the intended discrepancy
Weight can support a count only within limits that are easy to state and easy to skip. Before treating a mass reading as evidence of the intended shortage, the buyer needs a confirmed homogeneous component type, since a mixed bag defeats a single reference mass. Tare and packing condition have to be defined, because the bag and any insert add mass that is not the product. The reference piece mass and how much it varies matter just as much: if pieces vary, the total-mass window widens and a small difference can hide inside it. The scale's own resolution and repeatability have to suit the difference being judged. The method also needs physically counted validation states relevant to the intended discrepancy, checked under the same defined measurement and packing conditions.
If any required input is missing, the relevant validation is unavailable, or the readings for the compared states overlap, do not claim that the mass reading establishes the intended shortage detection. Use a direct count or another agreed, verified method. A conditional estimate can be written as:
estimated count = (gross mass − tare mass) ÷ reference piece mass
That expression is a conditional estimate, not a validation. It also says nothing about colour, wrong variant or card version, and those checks still have to happen separately. A mass match does not establish identity, and a mass difference does not explain itself. Whether weight can support an accurate count depends on the actual component and the conditions above, not on weight being inherently suitable or unsuitable.
Recording a mismatch
A record supports a discussion with the supplier. It does not by itself guarantee a replacement, a credit or any other entitlement; those depend on the agreed terms and the evidence.
| Field | Entry |
|---|---|
| Record / date | |
| Supplier | |
| Kit / variant / pack revision | |
| Lot / carton / module and checked-unit ID | |
| Component reference and position | |
| Expected count and unit | |
| Directly found count and identity | |
| Applicable colour / card / label finding | |
| Method used | |
| Weight inputs and validation reference, if used | |
| Observed difference | |
| Evidence / photo reference | |
| Recheck and result | |
| Buyer or inspection-team disposition | |
| Supplier response |
One hypothetical row shows how the fields connect:
| Field | Hypothetical example |
|---|---|
| Kit / variant / pack revision | Variant A, pack revision 02 |
| Component reference and position | Clip, ref A-01, inner bag |
| Expected count and unit | 20 pcs |
| Directly found count and identity | 18 pcs, correct variant and colour |
| Method used | Direct count |
| Observed difference | Short 2 pcs |
| Recheck and result | Recount confirms 18 pcs |
| Disposition | Hold the checked unit, request supplier investigation |
The fields are left blank on purpose in the first table. Fill them from the actual check, not from the template. The difference is recorded the way it has to be acted on: expected against directly observed, then classified by type. A missing piece and a wrong-variant piece are different findings and may share a cause; record them separately rather than folding both into one shortage count. Keep the affected checked unit identifiable and follow the agreed inspection plan to decide whether wider isolation or rechecking is needed. Keep the traceability reference tied to the check so the unit, the batch and the finding can be connected afterward. How the supplier wants that verified, and from which record, is something to agree before production rather than after a dispute.
Keep pack verification separate from lot acceptance
A passed check on the units that were opened is not an accepted lot. Verified contents on a checked unit, verified contents on a sampled set of units and accepted whole-lot quality are three different claims. Keeping them separate is what stops a clean count on twenty packs from being read as a guarantee about the shipment. The pack-level finding feeds the lot decision; it does not replace it.
Sourcing request
For kit sourcing, send the agreed component list, component references, pack format, variant information and quantities you need sourced to Scarecrow Garden Supplier Co., Ltd. Include your counting and inspection requirements in the RFQ so they can be discussed against the exact pack specification.