Every founder hears the same sentence: our minimum is ten thousand units. It sounds like gatekeeping. It is mostly arithmetic, and once you can do the arithmetic you can plan against it instead of arguing with it.

Where the number comes from

A production run has fixed costs that do not care how many units come off the line:

Changeover. The line has to be broken down, cleaned, swabbed and released before your batch and after it. That is most of a shift, whether you make two thousand units or thirty thousand.
Batch documentation. The same record gets written, checked and signed either way.
Release testing. Micro, viscosity, pH and fill weight are per batch, not per unit.
Component minimums. Your bottle mould, your cap, your carton: each has its own minimum from a different supplier, and yours is usually the highest of them.

Divide those fixed costs by the run size and you get the number that decides whether a factory says yes. At ten thousand units they spread thin enough to disappear into the unit price. Well below that they do not; they are still real, and they have to go somewhere. Usually into your cost per unit, which means into your margin, which means into a retail price that does not work.

Why ours does not move

We publish ten thousand units per SKU and we hold it, for a reason that is easier to say than to hear: a run that is too small to price properly is a run that hurts the brand buying it. You would get the units. You would also get a per-unit cost that makes the second order impossible, and we would both discover that six months later.

We would rather tell you in week one that the numbers do not support a run yet, and tell you what would change that, than take the order and let the arithmetic do it quietly.

That is a separate question from development. Bench work and the scale-up batch that precedes a production run are their own stage with their own scale: a scale-up batch runs on the real line, but it is not a full run. If you are still formulating, the minimum is not the thing standing between you and starting.

How to reach ten thousand sensibly

The number is per SKU, which is the part most launch plans get wrong. Four ways to make it land:

Fewer SKUs, more units each. Three shades at 3,300 each is three runs, three sets of components and three stability studies. One hero at 10,000 is one of each.
One fill size. A 30 ml and a 50 ml of the same serum are two component sets, two stability studies and two labels. Pick the one your customer actually buys twice.
Stock components. A stock bottle in a stock cap gets you to market at a fraction of the component minimum of a custom mould, and nobody has ever rejected a brand over a stock bottle.
Share a base. If your cleanser and your body wash can run off one surfactant base at different fragrance loads, that is one batch of chemistry, not two.

Launch narrow and deep. It is almost always better than launching a six-piece range you cannot reorder.

The question to ask a factory

Not what is your MOQ; everyone has a number and everyone will quote you the high one. Ask: what does my per-unit cost look like at ten thousand, at twenty-five thousand and at fifty? The shape of that curve tells you whether the factory has actually thought about your growth or is just protecting a line slot. And ask what happens on the reorder, because the second run is the one that decides whether the brand works.