Bag Factory Audit Questions: 12 Answers That Reveal a Reliable Partner

What an Audit Is Actually For

Most creators approach an audit as a verification exercise, as though the purpose were to confirm that a facility exists and has machines in it. That is worth knowing but it is the least useful part of the exercise, because a facility that does not exist is not the risk. The risk is a real facility that behaves badly when your order is late, when a material is unavailable, or when a defect rate comes back higher than anyone expected. Those are behavioural questions, and they are answered by how a partner responds to a specific question rather than by what a photograph shows.

The reframe that makes an audit useful is this: you are not checking whether the production side can make the bag, you are checking how decisions get made when the answer is inconvenient. A partner who tells you a material is unavailable three weeks before it matters is worth more than one who promises it will be fine. A partner who shows a defect rate from last month, including the bad weeks, is worth more than one who insists everything passes. Candour about problems is the strongest available signal about how problems will be handled.

There is a second purpose, and it is defensive in a useful way. An audit conducted in writing, with the answers recorded, becomes the reference for the programme. When the pre-shipment inspection standard is questioned in week seven, the answer is in the document. When a substitution is proposed, the material policy is in the document. When a schedule slips, the escalation path is in the document. The audit is the first version of the programme's constitution, and writing it down is the point.

Our Bag Production Partner for Campaign Creators page sets out where this sits in the wider selection process, and the production partner selection checklist covers the capability dimension that sits alongside it. This guide is the question list, and it is written from the creator's side of the table.

Before You Ask Anything: The Documents to Request

Ask for the document pack before the question list, because the pack answers about a third of the questions without anyone having to write a reply, and because the willingness to send it is itself informative. A partner who sends a pack within two working days is telling you something about how they will handle your requests during the run. A partner who takes two weeks, or who sends marketing material instead of documents, is telling you something else.

1. Business licence and scope. Confirms the registered entity and that its scope covers the goods.
2. Recent audit or verification report. A third-party facility audit, with its date, and any corrective actions noted.
3. Capacity statement. Line count, headcount, monthly output, and the current loading picture.
4. Recent inspection reports. Two or three actual reports from recent orders, with the defect data in them.
5. Test reports. Fabric, hardware and substance tests from a named laboratory, dated within a reasonable period.
6. Compliance certificates. Any standard the campaign needs, with its certificate number and validity dates.

The fourth item is the most revealing document in the pack and the one most rarely offered unprompted. An inspection report from a real order shows the defect classes found, the sample size, the accept or reject decision and what happened next. A portfolio of photographs shows none of that. Ask for reports from an order that had problems, if they will share one, and pay attention to whether the problems are described plainly.

The dates matter as much as the documents. A test report from four years ago tells you the material was tested once, not that the current lot meets the standard. A facility audit from before a change of ownership tells you very little. Ask for current documents and check the dates, and if a document is old, ask what has changed since. That question alone will separate a partner who maintains records from one who collects certificates.

One practical note about the pack: ask for it before you ask for a price. If the documents arrive first, the quote can be evaluated against known capability. If the quote arrives first, it tends to dominate the decision, and capability gets discovered later. The sequencing costs nothing and it changes which factor decides.

Questions 1 to 3: Capacity, Lines and Where Your Order Sits

The first three questions are about whether your order matters to the facility and where it will physically be made. They are not rude questions and a good partner will not treat them as rude, because they are the questions that determine whether your 35 to 50 day window is real or aspirational. What you are listening for is specificity: a named line, a stated loading, and a person who owns the schedule.

Question 1 is usually answered with a headline capacity figure, and the headline is the least useful part of the answer. What matters is the loading: a facility running at 95 per cent has no slack for your order and no slack for a problem, while one at 70 per cent can absorb a late material or a rework. Ask the loading question directly. A partner who answers it is one who knows their own schedule, and the production MOQ calculator helps you size what you need against what they have free.

Question 2 sounds like a detail and it is the one that decides whether your order is made by the team you met. Bag production is line-specific in a way that matters: the operators who sewed your approved sample are the ones who understand its difficult operations, and an order moved to a different line at short notice is an order being sewn by people seeing the pattern for the first time. Ask which line, and ask whether the same line will make the bulk.

Question 3 is the one creators skip and the one that matters most in month two. A named person with a stated response time is a commitment that can be tested, and it is testable immediately: send a question and see how long the answer takes. The response time you experience during the audit is the response time you will experience during the run, and no service level agreement written later will change it.

Questions 4 to 6: Materials, Traceability and Substitution

The second three questions are about what the bag is made of and what happens when the specified material cannot be bought. Substitution is the most common unplanned event in bag production, and the policy around it is the difference between a controlled change and a disaster. What you want is a rule: no substitution without written approval, and a stated warning period before it is needed.

Question 4 matters for a campaign in a specific way that general buyers do not experience. If a defect appears after delivery and 200 backers have the same issue, the only way to find out how many units are affected is to know which fabric lots went into which cartons. Lot traceability is not an audit nicety, it is the mechanism that turns a recall into a targeted fix. Ask what happens to the record after the cartons leave.

Question 5 should be answered with a process rather than with a reassurance. The useful answer describes who proposes the substitution, what evidence accompanies the proposal, how much warning is given and who signs it off. The worrying answer is any version of an equivalent will be used, because equivalent is a matter of opinion and opinions differ most at the moment the goods are already cut.

Question 6 is where the compliance position becomes real. A test report from a named laboratory, dated, for the current lot, with the test method stated, is evidence. A supplier's own declaration is a claim. A certificate from years ago is a souvenir. For a campaign selling into a market with substance limits, this question decides whether the goods can legally be sold, and it should be answered before the deposit rather than before the shipment.

Questions 7 to 9: Sampling, Pattern and Change Control

The third three questions are about how the bag gets defined and what happens when the definition changes. This is the block that most directly affects the 6 to 10 working day sampling window and the 12 to 15 working day window for a complex construction, and it is where a campaign's schedule is usually won or lost. What you are establishing is whether the sample you approve is the sample that governs the run.

Question 7 is about what physical and documentary objects change hands. A made-up unit is necessary but not sufficient: you also need the specification that describes it, because the specification is what the bulk run is made to and what the inspection measures against. A sample without a specification is an object that can be interpreted, and interpretation is where disputes come from.

Question 8 should be answered with a version number. Bag development runs through several revisions, and each revision changes something measurable: a panel dimension, a strap width, a hardware reference, a stitch density. If those changes live only in a message thread, then in week six nobody can say which version the bulk is being made to. A versioned revision list is one page and it prevents the most common cause of a wrong bulk run, which is the subject of the sampling communication questions every creator should also ask.

Question 9 is the one that protects the quality outcome, and it is worth being precise about where the retained unit lives. A golden sample held by the production side, referenced in the inspection instruction, is the standard that makes AQL 2.5 to ISO 2859-1 meaningful. Without a physical reference, inspection measures against a description, and a description is always more permissive than the object it describes.

Questions 10 to 12: Quality Data, Compliance and Escalation

The last three questions are about what happens at the end of the run and when things go wrong. They are the questions a creator is least likely to ask because they are uncomfortable, and they are the questions whose answers are most predictive. A partner who can describe their own failure handling in detail has handled failures before and learned from them. A partner who cannot has either never had one or has not learned.

Question 10 is the single best question in the list, and the reason is that it is unanswerable without either honesty or invention. Every production run has defects; the question is what they are and what was done. A partner who says the rate was 2.4 per cent last quarter, most of it stitch alignment, and we changed the jig, has given you three pieces of real information. A partner who says everything passes has given you none.

Question 11 is about the worst case, and it has a number attached that should be in the answer. A failed inspection means either rework or remake, and a remake costs the 35 to 50 day window again. Ask who pays for a remake caused by a production fault, how long it takes and whether the production slot is protected. That answer tells you what a failure actually costs, and it is the number to hold in reserve when you set the campaign delivery date.

Question 12 is the one that protects you during the run, and it should be answered with a name rather than with a process. A named escalation contact, reachable directly, with a stated response time, is worth more than any clause in a contract, because contracts are read after the event and people are contacted during it. Ask for the name, and ask what happens when that person is unavailable.

All Twelve Questions in Two Tables

The twelve questions in one place, in the order to ask them, with what each one is really testing. Read the third column rather than the first: the question is a means and the signal is the point. A partner can prepare an answer to any question; what they cannot prepare is the habit of answering with numbers, names and dates.

Questions 1 to 3 are answered fastest and they are the ones that establish the working relationship, so ask them first and observe how the answers arrive. If they come back within a day, with numbers and a name attached, the rest of the audit is likely to go the same way. If they come back after a week, with a brochure attached, you have learned the most important thing already.

Questions 7 to 12 take longer to answer because they require documents, and the time taken is part of the signal. A partner who produces a versioned revision list from an existing programme in a day has the system. A partner who takes a week to find out whether a golden sample is retained does not, and no amount of reassurance afterwards changes that. Record how long each answer took, not just what it said.

How to Read an Answer: Green, Amber and Red

An audit produces three kinds of answer, and the middle category is the one creators mishandle. A green answer is specific, dated, named and verifiable: a figure, a document, a person, a process. A red answer is evasive, generic, or contradicted by a document already supplied. An amber answer is a genuine partial response, given with a reason: we do not hold that record electronically, but here is the paper version, or that figure is not measured, but here is what we do measure.

• Green. Numbers, dates, names and documents, offered without being chased.
• Amber. A partial answer with an honest reason and a proposed alternative.
• Red. Marketing language, a refusal, or an answer contradicted by the documents.
• Amber is usually fine. Very few operations measure everything, and candour about a gap is a good sign.
• Red is rarely recoverable. One red is a question; three is a decision.

The fourth item is worth dwelling on because it is where audits go wrong. Creators who have read a checklist expect twelve green answers and treat anything else as failure, which pushes them towards partners who are good at performing an audit rather than good at running one. The reality is that a mid-sized operation will not have an answer to everything, and the partner who says so and offers an alternative is more useful than the one who claims completeness.

The fifth item is the decision rule that makes the whole exercise practical. One red answer is worth a follow-up question, because it may be a misunderstanding or a gap being covered. Three red answers is not a misunderstanding, it is a pattern, and the pattern will hold during your run. Score the audit, do not average it: a single red in the quality data question outweighs four greens elsewhere.

There is one more category worth naming, and it is the most dangerous because it looks green: the answer that is specific, confident and unverifiable. We inspect every order to AQL 2.5 is specific and confident, and it means nothing without a report. Ask for the artefact behind every claim. A partner who can produce the artefact is different from one who can produce the sentence.

Auditing Without Flying: The Live Walk and the Document Pack

Most creators will not fly to conduct an audit in person, and the good news is that an audit conducted remotely, properly, captures most of the value. What remote auditing loses is the sensory impression of a floor: whether it is organised, whether the line is busy, whether the schedule board is real. What it keeps is everything documented, and documentation is where the decisions actually get made.

1. Request the pack first, and score the speed and quality of the response.
2. Ask for a live video walk, unedited, at a time you choose rather than a time they choose.
3. Ask to see specific things during the walk: the schedule board, the material store, the inspection area.
4. Ask to meet the line contact, not only the sales contact, and ask a technical question.
5. Record the walk and write up the answers the same day, while the details are fresh.

The second item is the one that separates a real walk from a performance. A recorded walkthrough, or a walk scheduled at their convenience, can be staged. A live walk at a time you nominate, where you direct what the camera looks at, cannot be staged to the same degree. Ask to see the schedule board and read it. Ask to see the material store and the lot labels. Ask to see where inspection happens and what is on the table.

The fourth item is the highest-value minute in the whole audit. Ask the person who will actually run your line a technical question about your product: how the strap attachment would be reinforced, what stitch density the fabric needs, whether a particular pocket layout creates a bulky seam. A line contact who answers concretely is the person who will solve your problems. A line contact who defers entirely to sales tells you who is actually making decisions.

There is one thing a remote audit genuinely cannot establish, and it should be named rather than glossed over: whether the facility you are speaking to is the facility that will make your goods. Document checks, a live walk and a named line contact make this substantially more likely, but the definitive answer is a third-party inspection of the finished goods before they ship, held to AQL 2.5 to ISO 2859-1. That is the control, and it applies whether or not you flew.

What to Do With the Answers Before You Commit

The audit ends with a decision, and the decision should be written down in the same document. Record which questions were answered green, amber and red, what the follow-up actions are, and what you are accepting as a known risk. A campaign that commits without that summary has done an interesting exercise rather than a useful one, and the summary is what makes the difference.

Three specific items from the answers belong in the commercial terms rather than in the notes. The substitution policy, because it governs a change that will probably happen. The escalation contact, because it is who you call when it does. And the inspection standard with its acceptance criteria, because it is the only mechanism that converts a quality expectation into a decision about whether the goods ship. Those three should be written into the order, not just discussed.

Two items belong in your own campaign plan rather than in the contract. The remake duration, because it is the number that sets how much buffer your delivery promise needs, and the loading picture, because a heavily loaded facility has less room to recover a slip. Both are schedule inputs, and both are worth more to you than to the production side.

Finally, treat the audit as renewable rather than as a one-off. Capability, staffing and loading change, and an audit conducted eighteen months ago is a historical document. Re-run the twelve questions at the start of each campaign, or at least the six that concern capacity, materials and quality data. It takes an hour and it is the cheapest risk control in the programme.

Red Flags That Should End the Conversation

Some answers are not amber, they are terminal, and it is useful to name them so that a creator does not talk themselves out of an obvious conclusion. The most common reason creators ignore a red flag is that the price is good or the sample was nice, and both of those are things a well-run programme also offers. A good price is not a reason to accept a bad process.

• Refusal to share any inspection report, from any order, for any reason.
• No substitution policy, or a policy that reserves the right to substitute without notice.
• No retained golden sample, or no clear answer about what the bulk is measured against.
• No named line or coordinator, only a general sales address.
• Test reports that are years old, presented as if current, with no explanation.
• A defect rate that is claimed to be zero, which means it is not being measured.

The first and last items on that list are the two that matter most, and they share a logic. An operation that will not show its inspection data, or that claims to have none, is an operation that does not measure its own output. An operation that does not measure its output cannot improve it, and it cannot tell you the truth when something goes wrong, because it does not know.

The second item deserves a specific warning for campaign creators, because the consequence is asymmetric. A retail buyer who receives a substituted fabric can reject the goods. A campaign creator who receives a substituted fabric has 800 backers expecting a specific thing on a specific date, and rejecting the goods means a remake at 35 to 50 days. There is no cheap version of that failure, which is why the substitution answer is worth insisting on.

The right way to end a conversation is briefly and without hostility. Say what you needed, say it was not available, and leave. Production partners are not obliged to run an operation to your specification and you are not obliged to place an order, and the audit exists to find that out before money moves rather than after. Every campaign we have supported started with a version of these twelve questions, and the ones that went well started with good answers.

Frequently Asked Questions

People Also Ask