How Do You Audit a Reorder Process Before You Commit?

Buyers often ask for an audit when what they need is a review of one programme's reorder controls. Those are different exercises. A facility audit asks whether a site is capable and compliant; a process audit asks whether this style, on this programme, will come back the same next time. Both have value, and only the second predicts a reorder.

The second is also the one a buyer can do themselves, in an afternoon, from documents. It needs no travel and no certificate, and its findings are concrete: a register line that is loose, a reference that has faded, a change log that stopped six months ago.

What Does Auditing a Reorder Process Mean Here?

It means testing the six controls that make a repeat predictable, on the actual style, using the documents that exist today. It is not a site inspection, not a social-compliance review and not a capability study. Where a buyer needs those, they are separate exercises with separate reports.

The distinction is worth stating plainly because the two are often confused in a booking conversation. A compliance or social audit report tells a buyer whether a site meets a standard; it says nothing about whether run four of a particular style will match run one. Only the programme documents answer that, and they are usually available the same day.

There is also a timing point. A process review of this kind is worth doing before a programme commits to a year of reorders, because the findings are inexpensive to fix at that point and expensive to fix after two runs have shipped. A report produced afterwards is a post-mortem rather than a control.

Keeping the scope narrow is what makes the exercise useful. A broad audit produces a long report that nobody acts on. Six focused questions produce one page, and one page gets fixed before the next booking.

What Documents Should Be Produced on Request?

Six: the spec file with its revision log, the component register with retained references, the tolerance table, the change log for the last twelve months, the first-piece sheets from the last three runs, and the comparison sheets from the last three runs. Six documents, one request, sent before the visit.

How quickly and how completely the six documents arrive is itself a finding. A programme that sends all six within a day is a programme that keeps them. One that sends three and promises the rest is telling you something more useful than any line in the documents.

How Do You Test Whether the Register Is Real?

Pick three lines at random and ask for the retained sample. A register that names a brand, material, finish and code is a document; a register whose three random lines each have a physical sample in an envelope is a control. The samples are the part that is hard to fake and easy to check.

Then ask which lines have changed in the last year and why. A register that has never changed is suspicious, because components are discontinued continuously. Either the register is not being maintained, or changes are happening without going through it.

Ask one follow-up question that separates the two: which discontinued code was replaced most recently, and when was the buyer told. A maintained register produces a specific answer with a date. An unmaintained one produces a general assurance that changes are always communicated, which is the answer to a different question.

It is also worth asking how the register is kept current in the ordinary course rather than at audit time. The honest answer is usually that it is updated when a component moves, which is fine; the worrying answer is that it is reviewed annually, because discontinuation does not wait for the review.

How Do You Check the Retained Reference Chain?

Ask to see the working gold sample and the archive sample together, in the light they are compared under, and ask for the date of the last comparison between them. Three questions, and the answers tell you whether the reference is a standard or a souvenir.

Then ask what happens when they differ. A programme that has a written rule - replace the working sample from the archive, or agree a new reference from a current run - has thought about it. One that has not will argue about it the first time it matters.

What Does a Walk Through the Line Tell You?

Less than most buyers expect, and more than nothing. What it can show is whether the first-piece sheet is genuinely used, whether the retained reference is on the table at setting-up, and whether the operator can say what the tolerance is for the seam they are sewing. Those three observations are worth the walk.

What a walk cannot show is whether the last four runs matched. That is in the comparison sheets, not on the floor. Read the sheets first and use the walk to check that what the sheets describe is what actually happens.

How Is the Inspection Itself Audited?

Ask three questions about the inspection: which sampling plan was used and where the sample size came from; what the defect classification is and whether critical, major and minor are defined in writing; and whether the inspector works from the tolerance table or from experience.

The sampling question has a checkable answer. Under ISO 2859-1 at General Inspection Level II and AQL 2.5, a lot of 501 to 1,200 units takes a sample of 80 with an accept number of 5, and a lot of 1,201 to 3,200 takes 125 with 7. An inspection report that quotes those figures is reading the standard; one that quotes a round number is not.

One further check closes the loop: ask whether the inspector sees the gold sample and the tolerance table, or only the bag. An inspector working without the reference is grading from experience, and experience varies between people, between days and between styles, which is precisely what the reference exists to remove.

Ask for the inspector defect list as well. A written list of classified defects, with examples, is what turns an opinion into a repeatable judgement, and it is the quickest way to see whether two inspectors would grade the same bag the same way.

What Questions Expose a Paper-Only System?

Four questions do most of the work. Which change was the hardest to agree this year, and why? Which register line is most at risk of discontinuation? Show me a run that failed its comparison, and what happened. And who signs when the two sides disagree?

A live system answers all four from memory, with specifics, and usually with some irritation about a particular component. A paper-only system answers with documents, generalities, and no examples, because nothing has ever actually been tested by a failure.

How Are Findings Graded and Recorded?

Three grades are enough. A gap is something missing entirely - no register, no reference, no log. A weakness is something present but not maintained - a register with loose lines, a reference that has not been checked. An observation is something working that could be simpler.

Record strengths as well as gaps. A programme that knows which of its controls works will protect them when someone proposes a shortcut, which is the most common way a good system quietly degrades over a few seasons.

What Does a Good Programme Show on the Day?

A person can produce the six documents in an hour, can name the owner of the spec file, can show a comparison sheet with real readings on every run, and can describe the last change that was refused and why. Those four things are the profile of a programme that will come back the same.

It also looks slightly boring. There is no drama, no escalation, and no story about a crisis that was heroically solved. Predictability is unremarkable to the people running it, which is exactly why it is easy to dismantle by accident.

What Should Happen Before the Next Booking?

Close the gaps, note the weaknesses with an owner and a date, and put the review on the calendar. Nothing else. A process audit that produces a long improvement programme is a process audit that will be abandoned; one that produces six dated lines gets done.

Where a blocker was found - a gap affecting a tested or a regulated claim - the booking should wait. That is the one finding with a schedule consequence, and it is worth naming it as such at the time rather than discovering it when the goods are ready to ship and the claim cannot be supported.

Send the findings page to both sides, not just to the file. A gap understood by one side is a gap that gets quietly worked around; the same gap on both desks becomes a change to the way the next run is booked, which is the entire point of doing the review.

Frequently Asked Questions

People Also Ask