A sample and a production run are not the same process. The bulk drifts in four places: a different line, a different component batch, different hands, and changes nobody wrote down. Only the last one is dishonest.
The sample came back right. You held it, checked the button, looked at the colour in daylight, and gave the go-ahead.
Six weeks later the cartons open, and something is off. Not broken. Off. The colour is a shade deeper. The button feels loose. The logo sits two millimetres to the left.
Nothing was switched. Nobody set out to trick you.
What happened is simpler, and worse: you approved a sample, and a sample is not what a factory produces when it runs at speed.
A sample is a prototype, not a preview
A sample is built by one or two people, at their own pace, on a bench, with parts they picked up by hand.
Your order is built on a line, on a schedule, by whoever is on shift, with components from whatever batch is in the warehouse that week.
The sample took an hour. The order takes three days. Same drawings, same workers — and a completely different amount of attention per unit.
That's how a sample ends up flawless while the bulk disappoints, with nobody lying to you.
Four places it drifts
1. The line.
The sample was built on a sample bench. The order runs on a production line — different machine, different fixture, sometimes a different building.
Colour is where you notice it first. Two shades of the same Pantone look identical on their own and obvious side by side. Ink runs and dye lots are never identical, and at volume the difference compounds.
2. The components.
Same part number, different batch. That is not the same component.
A U-C connector from one supplier and a U-C connector from another carry the same part number and different tolerances. Cells come in a range, and not every unit gets the top of it. On a single unit, invisible. Across five thousand, a return rate.
3. The hands.
Every manual step has a tolerance, and that tolerance is whoever is doing it.
A glued seam, a trimmed edge, a soldered joint, a sticker placed by eye. The sample got the person who has done it for nine years. The bulk got three people, one of whom started this month.
4. The changes nobody logged.
Somewhere between your approval and the run, someone "improved" something. A thicker cable. An indicator light moved two millimetres. A capacitor swapped because the original was four weeks out.
Some of those are harmless. Some quietly change the product you approved. Either way, you hear about it when the cartons open — unless you asked before the line started.
What to freeze at the sample stage
This is the cheapest hour in the whole order. And before you sign anything, it's worth knowing what you're actually looking at — what to check on a sample before you approve it.
- Sign the sample, not a photo. Date it, both parties, keep one each.
- Define the golden sample in writing. A golden sample is the unit both sides signed — the physical object your order gets measured against. If it lives in a drawer on the line, it will drift within a month.
- Write down what's allowed to vary. Colour tolerance, weight, capacity, dimensions. Anything you didn't write down is open to negotiation at 2am on the line.
- Ask one question: was this sample built on the sample bench, or on the production line? If it was the bench, ask what changes when the order runs.
- Ask for the first ten. Not a QC report — a photo of the first ten units off the line, next to your signed sample, before the rest is packed.
What it costs to fix later
Once the container lands, you have three options, and none of them are cheap.
Sort the batch — labour, plus a delayed listing.
Ship it back — rarely worth it.
Sell it and note the difference in the listing — until the reviews arrive.
That's why the sample hour is worth more than it looks. It's the only moment when the fix is free.
And when it isn't free — when the container is already in your warehouse — who actually pays for a wrong first batch is a separate question with its own answer.
Straight answers
Why does the bulk look different from the sample, if it's the same factory?
Because the sample and the run are two different processes. One person at a bench is not a line running on a schedule. The factory didn't change. The method did.
The bulk is close but not identical. Is that a defect?
Only if it falls outside what you wrote down. Which is the whole reason to write it down. "Close" without a number is an argument waiting to happen.
Who pays for a second sample, or a re-sample?
Usually you, unless the first one missed the spec you both signed. Ask before you order, not after.
The honest part
A factory doesn't send you a good sample to deceive you. The sample is what they're capable of on a good day.
Consistency is a different number, and it's a property of the line, not of the sample. You can't inspect it in. You can only pay for it — in time spent on the line before the run, or by choosing a factory that has made this exact part before.
The sample tells you what's possible. The first ten units tell you what you're actually buying.
— Damon, Founder
