What Numbers Belong in a Quarterly Supplier Inspection Review?

A quarterly review is the one governance meeting that most buyers hold and few prepare. An hour is booked, the agenda is a list of topics rather than a list of numbers, the conversation runs in adjectives, and the meeting ends with a general sense that things are going well. Nothing is decided, because nothing measurable was placed in front of the room, and the next quarter repeats the same meeting with the same result.

The fix is not a longer meeting. It is a rule about what is allowed onto the agenda: an item appears only if a number can be set beside it, and that number has to come from inspection output, delivery records or the corrective action log. The rule turns the review into a decision meeting, and it has a second effect worth as much. Once everybody knows where the quarterly data ends up, the data gets collected properly during the quarter instead of being reconstructed in the week before the meeting.

What Numbers Come Into a Quarterly Review?

Five sets, and the discipline is to keep it to five. Inspection outcomes by lot, expressed as accept, accept after rework and reject. The defect class split inside those lots, counted against the agreed class list. Delivery against the window agreed when the slot was booked, lot by lot. Sampling rounds, meaning how many rounds each style needed before written approval. And the corrective action log: items opened, items closed and items still open.

Everything else is context rather than data, and it should be handled as context. Correspondence about one late shipment is evidence about one delivery, not a metric. A complaint from a retail customer is an input to the defect class list rather than a line on the agenda. A general observation that communication has improved is worth hearing once and is worth nothing if it is recorded as though it were measured.

How Is the Accept and Reject Outcome Counted by Lot?

By lot, and never by unit. A lot is the quantity presented for inspection at one time, and each lot carries one outcome. A quarter of six lots with one reject is a reject rate of one in six, not one in thirty thousand units. Unit counts hide the pattern, because a single large lot dominates the arithmetic and a small lot that failed twice disappears inside it.

Three outcomes rather than two. Accept, accept after rework, and reject. The middle category is the one most often dropped and the one that carries the most information: a lot released only after rework consumed capacity that was not planned for, and a rising count of them is the earliest signal that a specification, a material or a station has drifted. Record the rework days next to it, because that number is what the next quarter loses.

How Should the Defect Class Split Be Read?

As a distribution rather than a total. Critical findings should be zero, and any critical finding is an event rather than a statistic: it gets its own line, its own cause and its own corrective action item, because at AQL 2.5 a single critical defect fails the lot regardless of how the rest of the sample performed. Major counts drive the accept or reject decision, so a rising major count is the number that predicts next quarter reject.

Minor counts are the ones that erode a programme quietly, because they sit below the threshold that fails a lot and above the level a retail customer tolerates. Read them as a trend rather than as a verdict, and read the whole split against the same class list all quarter. Where the class boundaries move between lots, the split stops being comparable and the exercise collapses into argument about definitions.

How Is Delivery Measured Against the Agreed Window?

Against the date agreed when the slot was booked, not against the date the buyer hoped for. On this programme bulk production runs 35-50 days from written sample approval and sampling runs 6-10 working days from a signed-off tech pack, so the agreed date is a calculated date rather than an aspiration. The metric is goods-ready minus agreed, in days, per lot, with the sign kept.

Report the spread as well as the average. A programme that averages two days late but ranges from ten days early to fifteen days late has a planning problem that the average conceals, and the spread is what the buyer actually has to hold stock against. Where the cause is known, split the lateness by cause: buyer approval delay, material lead time, capacity, or inspection rework. That split decides who owns the fix, and without it every late lot becomes an argument.

Why Do Sampling Round Counts Belong on the Agenda?

Because the number of rounds a style needs before written approval is the cleanest measure of how well the specification is understood. One round means the tech pack was readable and the construction was within the normal capability of the line. Two rounds is normal. Three or more means something in the pack was ambiguous, and the ambiguity is sitting in a document the buyer owns.

It is also the number with the largest effect on the calendar. Each additional round costs a full cycle of 6-10 working days plus the buyer approval time that follows it, and that cost lands directly on the launch date. A style needing three rounds reaches bulk roughly a month later than the same style needing one, and the quarter review is the only place where that pattern becomes visible across styles rather than one style at a time.

How Are Corrective Action Items Counted as Opened and Closed?

Three counts and one age. Items opened in the quarter, items closed in the quarter, items still open at the end of it, and the age of the open set. The first two show the flow, the third shows the backlog, and the fourth shows whether the backlog is moving or simply ageing in place. A programme that opens twelve and closes eleven looks healthy; a programme whose open set has a median age of two quarters does not.

Closure has to mean something countable. The position that works is that an item closes when the correction has been verified against evidence: a revised document, a re-inspected lot, a photograph of the changed station, a test report. An item closed on a written statement alone reopens in practice, usually at the next inspection, and the log then records a closure that never happened.

What Is the Right Agenda Order?

Work from the record towards the decision, and put anything that needs a commercial choice last. Seven items in a fixed order: inspection outcomes, defect class split, delivery, sampling rounds, the corrective action log, the forward plan, then actions and owners. The order matters because the first five produce the facts that the last two act on, and a review that opens with the forward plan spends its hour on a plan built from nothing.

Who Attends, and What Does Each Person Bring?

Four roles, and the rule is that each brings a document rather than a view. The quality lead brings the inspection reports and the defect counts. The technical owner brings the tech pack versions and the sampling history. The programme owner brings the delivery record and the corrective action log. The commercial owner brings the forward volumes and the launch dates.

Two attendance rules are worth writing down. Somebody who can commit a date must be present, or the forward plan item produces a range instead of a plan. And the meeting should have a chair who is not the person with the most at stake in the outcome, because the chair is the one who decides whether an item is closed or carried, and that decision has to be seen as neutral on both sides.

What Gets Recorded During the Meeting?

One page per quarter, circulated within a day. The five numbers with the previous quarter beside them, the exceptions discussed, the corrective action items with owners and dates, and the decisions taken. One page is enough, and one page is what gets read. A full minute is written and then never opened again, and the following quarter then starts from recollection.

Record what was decided not to do as well. A decision to leave the inspection level unchanged, or to accept a delivery pattern rather than rebuild the window, is a decision that will be questioned later by somebody who was not in the room. One line stating the decision and its reason converts a future question into a lookup.

Where the governance file holds a social compliance report, the review handles it as evidence rather than as a judgement. The report date, the report validity and the open findings are recorded, and the findings themselves are assessed by a qualified third party against the law of the production country and against the buyer code of conduct. A quarterly review is not the place where a compliance position is decided, and it should not read as though it were.

What Comes Out of the Meeting?

Four things, and a meeting that produces fewer than four has not finished. A decision on each agenda item, including the decision to change nothing. A set of corrective action items, each with an owner and a date. A revised forward plan with slot dates and a freeze date. And a change record wherever something that was frozen has moved.

The change record is the one most often skipped and the one that matters most over a year. A specification that drifts through four small changes between reviews is a specification nobody can build to consistently, and the drift is invisible until somebody writes each change down with a date next to it. The record takes five minutes and it is the artefact that makes a reorder a reorder.

How Does the Review Connect to the Scorecard and the Contract?

The review is where the scorecard is fed. The five numbers are the scorecard inputs, and the scorecard is what makes them comparable across quarters and across production partners. Holding the review without the scorecard produces a series of conversations; holding the scorecard without the review produces a number nobody has discussed. The two instruments only work together.

The review is also where the contract facts get checked rather than assumed. Is the inspection level in force the one agreed. Does the window quoted on the last order match the window the programme was planned against. Are the payment terms being applied as written. Those three checks take five minutes, and they catch the quiet drift that turns into a dispute in the second year.

1. Inspection outcomes

Counts by lot, with the previous quarter beside them

2. Defect class split

Critical, major and minor summed across the quarter

3. Delivery

Goods-ready against agreed date, with the spread and the cause

4. Sampling rounds

Rounds per style, first submission to written approval

5. Corrective action

Opened, closed, open, and the age of the open set

6. Decisions

One line each, including the decisions to change nothing

7. Actions

Owner, action and date, agreed before the meeting closes

What If the Review Finds Nothing to Discuss?

Then the review is working, and the right response is to shorten it rather than to cancel it. A quarter with no exceptions is evidence that the specification is stable, the class list is doing its job and the planning window is realistic. Those are worth fifteen minutes of confirmation and one page of record, because they are also the conditions that make the next quarter predictable.

Keep the record anyway. The value of an unbroken series of quarters is that a change in the numbers is visible when it arrives, and a series with gaps cannot show a change. The programmes that manage drift well are the ones with three years of one-page records sitting in a folder, and the ones that manage it badly are the ones that started recording the quarter after something went wrong.

The commercial frame is what the review measures against: MOQ 500 pcs per style and colourway, sampling 6-10 working days from a signed-off tech pack, bulk production 35-50 days from written sample approval, inspection at AQL 2.5 under ISO 2859-1 General Inspection Level II, T/T 30/70, quoted FOB Xiamen with CIF and DDP available. A delivery number is only meaningful when the window behind it is one of those figures.

One further point about pace and about honesty. A review where the buyer side has not done its own work — approvals late, forecasts missing, artwork supplied after the sampling round started — will read as a production problem in the numbers. The agenda should carry a line for what waited on the buyer, and that line is as useful in quarter one as it is in quarter twelve. Most delivery drift in a young programme is buyer-side, and most buyer-side drift is invisible until it is written down.

Frequently Asked Questions

People Also Ask