Nobody can give you the numbers for your project, and anyone who gives them to you without having seen your product is selling something. What can be given is the structure: which lines exist, which of them recur, which ones move when the project changes shape, and in what form to present the whole thing to someone who has never worked with China. A budget built that way survives contact with reality, because it was never a single number.
The lines that get left out
The forgotten lines are not exotic. They are the ones that sit between the decision and the first sale, which is exactly the period a budget written at the decision tends to skip. Each one is predictable; the reason it is missing is that it belongs to no single department and appears on no quotation.
| Line | One-off or recurring | What makes it move |
|---|---|---|
| Travel and time on site | Recurring, and more often than planned in the first year | The number of suppliers or partners you are still choosing between, and whether anyone local represents you |
| Samples and the rounds they take | One-off per development, but repeated at every modification | How complete the specification was at the start — an incomplete spec is paid for in sample rounds and air freight |
| Tooling and fixtures | One-off, and payable long before any revenue | The number of variants, and who ends up owning the tool — a point that belongs in the contract, not in the budget |
| Translation and bilingual documents | Both: a one-off for the first set, recurring for every update | Whether the document is commercial, technical or legal — three different trades at three different levels of care |
| Certification, testing and registration | One-off per product and per market, recurring when the product or the rule changes | The markets you actually name. « Europe » is not a market and the same product can need different files in different countries |
| Internal time | Recurring, and the largest line nobody writes down | The time zone gap, the number of people involved, and whether one person owns the project or three share it |
A one-off and a recurring cost are two different decisions
A one-off cost is a bet: you pay it, it is gone, and what it buys either works or does not. Tooling, a registration, a first trip, a trademark filing. Its risk is concentrated at the moment of payment, and the question to ask about it is what you still own if the project stops — a registered trademark survives the project, a sample round does not.
A recurring cost is a commitment, and it is the one that actually decides whether the project reaches its second year. It compounds quietly, it rarely appears in the business case, and it is almost always under-estimated because the first months are the cheapest. The useful discipline is to write recurring lines as a rate — per month, per trip, per order — and to multiply by the duration you genuinely expect rather than the one you hope for. A project budgeted as a one-off and lived as a recurring cost is the standard way a sound idea runs out of money.
Ranges, not numbers
- Give every line a low and a high. A single figure claims a precision you do not have, and the person reading it will treat it as a commitment. A range says what you know, which is the shape of the cost rather than its value.
- Say what each range depends on. « Low if we keep one variant, high if we keep three » turns a budget into a decision document. A range with no stated driver is just a wider guess.
- Separate what is quoted from what is estimated. Two columns, or two colours. A line backed by a written quotation and a line you invented deserve different levels of trust, and the reader cannot tell them apart unless you say so.
- Put the contingency where it belongs. A single global cushion at the bottom gets cut in the first review. A contingency attached to the lines that are genuinely uncertain — development, certification, the first production run — survives, because it is argued.
- Budget the currency and the payment moment, not just the amount. A cost agreed in one currency and paid months later in another is two decisions, and the second one belongs to whoever manages your cash rather than to the project.
When to revise it
A budget revised continuously is noise; a budget never revised is fiction. The workable rhythm is to revise at events rather than on a calendar — when the specification is frozen, when the supplier is chosen, when the first quotation arrives in writing, when the first samples are approved, when the first order ships, and when any market is added. Each of those replaces estimates with quoted figures, which is the only kind of revision worth making. Between events, record what you actually spent against each line, because the gap between the two is the information you will need for the second year — and it is also the only thing that makes your next budget credible.
Presenting it to someone who has never worked with China
The reader who has never done this has one honest question behind every other: what am I committing to, and when can I stop? Answer it structurally. Split the budget into stages that each end in a decision, with the money needed for that stage and nothing beyond it. Explain, for each stage, what you will know at the end that you do not know now. A budget presented as a single total invites a yes or a no; the same budget presented as three gates invites a yes to the first gate, which is all you need and all you can honestly ask for. Two habits help more than any formatting. Say out loud which line you are least sure of, before anyone asks — it buys you the credibility to be believed on the others. And give the cost of not doing it alongside the cost of doing it: a competitor's landed price, a supply that depends on one source, a market you are not in. A budget that only shows spending looks like a risk; a budget that shows a choice looks like a decision.
Keep the budget and the actual spend in the same file, line by line, from the first month. It takes minutes and it is the only document that will let you argue for the second year — or decide, with something in your hand, that there should not be one.