Initial Production Check: The Cheapest Fix in the Order

An initial production check exists because the same mistake costs one afternoon at unit fifty and one container at unit five thousand. It is the earliest point where you can change what the factory makes, instead of judging what it already made.

Everything in the right column is the same problem as the left, found later and at full quantity.

The First 10% to 20% Is the Whole Point

The check is booked once regular production has started but most of the order is still unmade, often somewhere in the first 10% to 20% of the run. Any earlier and there is nothing to look at. Any later and the mistake is already built into much of the order, which turns a correction into a negotiation.

That window matters because it is the last point where the factory can change course cheaply. Materials can still be swapped. Workers can still be shown the right way. A machine setting can still be changed before it is built into every unit.

This is the first check to add when something is new, and the first to drop when nothing is. That trade-off is a quality control in China decision, and it gets made order by order rather than once.

Being on the floor inside that window, rather than whenever the factory finds it convenient, is manufacturing control work.

Its answer is not simply yes or no; it tells the factory what to change while the bulk of the run is still ahead. A pre-shipment inspection tells you whether to accept goods that already exist. This one lands while there is still something to change.

The Sample in the Office, and the One at the Workstation

The most common thing an early check finds is not a defect but the wrong reference. The approved sample sits in the sales office, signed and tagged, while the line is working from last season’s version, or a photo on a phone, or what the supervisor remembers from the last buyer who ordered something similar.

Nobody is lying; the information simply never walked from the office to the floor. The person who agreed your spec is not the person building it, and in most factories those two are separated by a sales team, a production manager, and a shift handover.

This is why so many good samples fail at prototype to mass production. The sample was right. The order was right. The floor was working from something else, and the first time anyone compared the two was after the goods were packed.

So the first question on day one is not whether the units look good. It is what the line is comparing them to. Ask the inspector to photograph what the floor is actually working from: the approved sample, the drawing, the color chip, the label proof. If it is not physically at the workstation, you have found your defect before a single unit is checked.

What Day One Actually Covers

Start with the materials, because a wrong input cannot be inspected out later. Grade, thickness, color, finish, and whether the batch on the floor is the one the factory claims it bought.

Move to the parts, because the wrong one stops everything. The plug type, the fastener, the label, the retail card, the carton. A missing part delays production, but a wrong part that still fits spreads through the batch unnoticed.

Finish with the units already built, checked against the reference you just verified. Fit, finish, markings, packing. They are where you finally see what the factory thinks you ordered.

What it does not cover is the whole order, and pretending otherwise is how buyers get burned. The first 10% to 20% shows whether the line started with the right materials, settings, and reference. It does not tell you the last carton will be right. That is what a during production inspection later in the run is for.

The Report Needs a Decision, Not a Discussion

A useful report ends in one of three actions: keep going, fix the process, or stop the affected work. Anything in between is how a small problem becomes a shipped one. A report that lists problems while the same work carries on unchanged buys you nothing but a dated note of when you noticed.

So decide the rule before the inspector arrives, not while a factory manager is on the phone. Write it into the purchase order: a critical fault, or a major one that keeps repeating, pauses the affected work until you clear it in writing.

Then be willing to use it, because the pressure runs the other way. Stopping a line costs the factory money today and costs you a few days of schedule. Not stopping it costs whatever the mistake is worth multiplied by your order quantity, and you find out at the end, when it is a failed inspection in China instead of a production issue.

When It Is Worth the Day Rate, and When It Is Not

Book it when the order is new, the spec is fussy, or the factory is unproven. First order at a new supplier, a custom or private-label product, tight colors, retail packaging, anything where close enough comes back as a return.

Skip it when you are reordering a simple product from a factory that has already built it correctly, more than once. The China inspection cost of a day rate on a repeat run of plain goods is real money for thin margins.

Book it if anything changed. A new material source, a new factory building, a long gap since the last order, or a spec update you sent by email and never confirmed. Factories change quietly, and the first units are where the change shows.

New product, new factory, new material, or a specification the factory has never built to: any one of those earns the early check. A fourth repeat of the same item from the same line is where you can drop it, and that is the only case where dropping it is cheap.

FAQ

Q1: Can the factory just send photos of the first units instead?

Photos show you what they chose to photograph. They cannot show you which reference the line is using, what the material batch is, or what the units nobody selected look like.

Q2: What if production has not really started on the booked day?

You may still be charged if the booking was confirmed and the delay surfaced too late to cancel, so use the day: materials, parts, and the reference the line will work from. Then ask for a revised schedule based on the real start date, because a late start rarely stays a small delay.

Q3: What if only one of several SKUs has started?

Check the one that is running, then ask when the others start and whether they use the same materials, line, and workers. One good SKU proves nothing about the versions built next week by a different team.

Q4: What should I do when the material labels do not match the purchase order?

Do not settle for a verbal explanation. Ask for the batch numbers, the supplier records, and any test reports that tie the material on the floor to the one you approved, and treat anything the factory cannot trace as unverified.

Q5: Is the factory right that early units are always rough?

Sometimes, for the finish on the first pieces off a new setup. But that is not an excuse for wrong materials, parts, sizes, or markings, and a factory that blames the warm-up for those is telling you your spec never reached the floor.

Q6: Can production continue while small problems are fixed?

Yes, when the fault is clearly contained and cannot reach the remaining units. Pause the affected work whenever the cause is still unknown, because an unexplained defect is one that repeats.

Q7: Who has the authority to stop the line?

You do, and only if you wrote it down. The inspector reports; the purchase order decides whether that report pauses production or just gets filed.

Q8: Should I pay for a second visit before the run continues after problems are found?

For anything serious, or likely to repeat across the batch, usually yes. A fix nobody checks is just a promise with a photo attached. Wait too long and the whole order is finished before anyone looks again.

Conclusion

Later inspections still catch problems; this one changes the process before the mistake spreads through the rest of the run. That is the whole argument for a day rate at the front of a run, and it is why the buyers who skip it usually pay the same money later with freight attached.

Standing on the floor at unit fifty costs one day rate, and reading the same news at the end of the run costs the order. We take the early slot through manufacturing control, so the factory changes course while changing course is still free.