# What a QC Record Should Contain | Cheery Supply Chain

> What a tackle factory's QC paperwork should contain, what each field lets you check months later, and what a pass does not prove.

Source: https://luressource.com/insights/what-a-factory-qc-record-should-contain/

Direct answer. A QC record is worth filing only if it names the batch, the sampling rule, the method, the measured result and the disposition. Most inspection sheets a buyer receives record a verdict and nothing else, which makes them impossible to audit and impossible to compare between orders. Ask for the batch link, the written sampling plan, the measured values rather than a tick, the defect classification, and the name of whoever released the goods. Key facts. 1) A measured value with a batch number can be re-checked months later; a tick in a box cannot. 2) "AQL 2.5" is one parameter, not a sampling plan. 3) Re-inspection after a fix does not erase the original non-conformance. 4) Records that stop at the factory gate cannot support a claim after the goods arrive. Frequently asked questions. Q: Is a third-party inspection report better than the factory's own record? A: They answer different questions. A third-party report gives you an independent verdict on one shipment; the factory's own records are what let you see a trend across shipments. Buyers who only ever commission third-party inspections end up paying for the same discovery every order. Q: How many units should be inspected? A: That is set by the sampling plan, not by preference. Once the lot definition, inspection level and acceptance limits are written down, the sample size follows from the standard. The important thing is that all three are agreed before production, not chosen when the goods are ready. Q: The factory says it has ISO 9001. Is that enough? A: A certified management system says the factory has a documented process, not that your order was inspected to a rule you agreed. Ask for the record for your batch. If the system is real, the record exists; if it does not exist, the certificate is telling you about the factory's paperwork rather than your goods. Q: What if the factory will not share measured values? A: Treat that as information. Some factories genuinely do not record values and only tick boxes; others record them and prefer not to share. Either way, the fix is the same: put the reporting requirement in the purchase order before you place it, and make the record a condition of the final payment. Q: How long should records be kept? A: Long enough to cover the claim window you agreed. If your terms allow a defect claim 90 days after arrival, the record has to survive at least that long and be retrievable by batch. Ask where it is stored and who can pull it. Sources (9). 1) Attribute sampling framework referenced above — ISO standards catalogue — https://www.iso.org/ 2) Sampling terminology used on this site — our glossary entry for AQL — /glossary/aql/ 3) Surface and corrosion test terminology — our glossary entry for the salt spray test — /glossary/salt-spray-test/ 4) How we check a supplier before an order — our published verification method — /verification/ 5) ISO standards catalogue — quality and testing frameworks — https://www.iso.org/ 6) Chinese business registration lookup — National Enterprise Credit Information Publicity System — https://www.gsxt.gov.cn/ 7) Chemical compliance for plastics and metal finishes — ECHA (EU REACH) — https://echa.europa.eu/ 8) Harmonized System heading 9507 (rods, hooks, reels, lures) — US HTS lookup — https://hts.usitc.gov/ 9) US consumer product safety requirements — CPSC — https://www.cpsc.gov/ An inspection report tells you what somebody decided. It rarely tells you what they measured, on which batch, by which rule, or who signed it off. Those four gaps are why two buyers can hold two certificates for the same product and still have no idea whether the goods match. This page sets out the fields a QC record needs before it is worth filing, and the questions that turn a certificate into evidence. It is written for buyers placing a first or second programme, and for anyone who has just received a PDF of ticked boxes and is trying to work out whether it means anything. The five fields that make a record auditable A record is auditable when somebody who was not there can reconstruct what happened. Five fields do that. Everything else is decoration. Field What it has to say What its absence costs you Batch or lot link Which production run this record covers, tied to the order and the SKU version You cannot tell whether the sample came from your order or from last month's stock Sampling rule How many pieces were drawn, from where in the run, and what acceptance limit applied A 3-of-3 pass and a 315-of-315 pass look identical on a certificate Method The written procedure, its version, and the equipment used Nobody can repeat the test or challenge the result Result Measured values and defect counts, not a verdict You cannot see a trend forming across three orders Disposition Who released the goods, when, and what was done with the rejects Failed units can reappear in the next shipment with no trace Notice what is missing from that list: the word "pass". A verdict is the conclusion. The five fields are the argument. Buyers who collect conclusions end up arguing about opinions; buyers who collect arguments can settle a dispute with a photocopy. "AQL 2.5" is not a sampling plan Most inspection sheets a buyer receives say something like Inspected to AQL 2.5 . That phrase is a fragment. The acceptance quality limit is one parameter inside a sampling scheme, and on its own it does not tell you how many units were opened. The sampling framework most factories and third-party inspectors work to is ISO 2859-1, the attribute-sampling standard that grew out of the older MIL-STD-105 family. Using it properly requires three decisions to be written down together: The lot — what counts as one batch, and therefore what the sample is drawn from. The inspection level — normally one of the general levels, which sets how the sample size scales with the lot size. The acceptance limit, split by defect class — critical, major and minor are usually given different limits, and a single number collapses that distinction. Change any one of the three and the number of units inspected changes, sometimes dramatically. That is why "AQL 2.5" on a certificate is not evidence of anything except that somebody wrote it down. A usable plan reads more like: one lot per SKU per shipment; general inspection level II; critical 0, major 1.5, minor 2.5; sample drawn after packing, from at least three cartons. We are not telling you which numbers to use. We are telling you that the numbers have to be agreed in writing before production, appear on the purchase order, and appear again on the record. If they only appear on the record, they were chosen after the goods were made. Three records, not one Factories often hand over a single final-inspection sheet and call it QC. It is one of three records that do different jobs. Incoming material This is where most quality is actually decided. For a soft lure the incoming checks are the resin or compound batch and its hardness; for a metal jig it is the lead or zinc alloy and the hook lot; for a rod it is the blank and the guide set. The record should name the supplier batch and the receiving date. Without it, a defect found three weeks later cannot be traced to a material lot, and the factory can only guess at the cause. In-process In-process records catch the failures that final inspection cannot afford to. Colour drift during a long injection run is the classic example: the first hundred pieces match the approved colour chip and the last hundred do not, and a final sample drawn from the top of the carton may miss it entirely. Weight checks on jigs, hardness or sprue checks on soft plastics, and guide alignment on rods all belong here. Final and pre-shipment Final inspection answers one question: may these goods be packed and shipped. It should be recorded after packing, not before, because packing introduces its own failures — wrong label, wrong assortment, hooks not protected, cartons under-spec. A record dated before packing cannot speak to any of that. The defect classification argument Almost every dispute about a QC result is really a dispute about classification. A hook point that is slightly rolled is a minor defect to the factory and a major one to the buyer, because the buyer knows it will come back as a return. Classification is therefore not a technical exercise, it is a commercial one, and it has to be settled while the goods are still theoretical. The workable approach is to write three short lists into the specification: Critical — anything that makes the product unsafe or unusable, plus anything that breaks a regulated claim. Zero acceptance, no negotiation. Major — anything a customer would return. This is the list worth spending time on, because it is where the two sides genuinely disagree. Minor — cosmetic deviations within an agreed tolerance. Attach photographs with the lists. A written definition of "paint run at the eye" is worth less than one annotated photo, and it costs nothing to produce at the sample stage. What "passed" does not mean Four things get read into a pass that the pass does not say. It does not mean every unit was checked. It means a sample met a rule. That is the whole claim. If the rule was never written down, the claim shrinks further. It does not mean the batch is uniform. Sampling assumes the lot is broadly homogeneous. Injection moulding runs, plating baths and hand-assembly lines all produce drift. A pass on a sample drawn from one carton tells you about that carton. It does not survive a change. Swap the hook supplier, change the compound, move the mould to a second machine, or change the packer, and the previous record describes a product that no longer exists. This is what the sample-to-production gap means in practice, and it is the single most common reason a good sample is followed by a disappointing order. It does not travel. A record that stops at the factory gate cannot support a claim after the goods arrive. If the terms you agreed allow you to raise defects 60 or 90 days after delivery, the record has to be retrievable for at least that long. Ask now, not after the first complaint. Re-inspection, and why the original record stays When a defect is found, the usual sequence is: the factory sorts or reworks, re-inspects, and sends a clean sheet. Two habits make this safe rather than merely reassuring. First, keep the failing record. A re-inspection that supersedes the original is a re-inspection that has erased its own history. The useful file has both, plus a note of what was done with the rejected units. Sorted-out rejects have a way of reappearing in the next lot, and a record is the only thing that catches it. Second, make the fix specific. "Reworked" is not a corrective action. Retightened the hook eye on 240 units, replaced the split rings on 62, repacked 18 cartons because the label was on the wrong face — those are actions, and they can be verified. The same discipline applies when a sample itself fails, which is covered in what to do when a sample fails . Put the requirement in the purchase order None of this works as a conversation. A reporting requirement that only exists in an email thread will be honoured on the order where somebody remembers and quietly dropped on the next one. The place to fix it is the document both sides sign. Five lines are usually enough. The record to be provided: incoming, in-process and final, or a named subset. The sampling plan: lot definition, inspection level, acceptance limits by defect class. The form of the result: measured values and defect counts, not a verdict alone. The release signature: who at the factory is authorised to release the shipment. The retention period, and the condition it attaches to — typically release of the balance payment or the start of the next order. Factories that already run a disciplined system will produce all of this without much trouble, because they are doing it anyway. Factories that do not will push back, and the pushback is useful information about what you are buying. It is better to learn that at the quotation stage than at the port. Where this goes wrong in practice The failure is rarely a factory refusing to inspect. It is a buyer accepting a document that looks like a record and is not one. A one-page certificate with a green tick, a date and a logo will pass through most internal approvals, because it satisfies the process without answering the question. Two habits close the gap. Ask for one measured value — the actual weight in grams on ten jigs, the actual hardness reading on ten soft baits — and see whether it arrives. Then file the record against the batch, so that when the next order arrives you are comparing two documents rather than reacting to one. Related reading: AQL , salt spray test , how to set a colour standard , and what to do when a sample fails . Frequently asked questions Is a third-party inspection report better than the factory's own record? They answer different questions. A third-party report gives you an independent verdict on one shipment; the factory's own records are what let you see a trend across shipments. Buyers who only ever commission third-party inspections end up paying for the same discovery every order. How many units should be inspected? That is set by the sampling plan, not by preference. Once the lot definition, inspection level and acceptance limits are written down, the sample size follows from the standard. The important thing is that all three are agreed before production, not chosen when the goods are ready. The factory says it has ISO 9001. Is that enough? A certified management system says the factory has a documented process, not that your order was inspected to a rule you agreed. Ask for the record for your batch. If the system is real, the record exists; if it does not exist, the certificate is telling you about the factory's paperwork rather than your goods. What if the factory will not share measured values? Treat that as information. Some factories genuinely do not record values and only tick boxes; others record them and prefer not to share. Either way, the fix is the same: put the reporting requirement in the purchase order before you place it, and make the record a condition of the final payment. How long should records be kept? Long enough to cover the claim window you agreed. If your terms allow a defect claim 90 days after arrival, the record has to survive at least that long and be retrievable by batch. Ask where it is stored and who can pull it.