Um banco lê primeiro os números e para ali se não fecharem. Por isso a projeção é calculada, não escrita.
Calculado a partir do que você inseriu, não escrito por nada. É a parte lida primeiro.
| Ano | Receita | Lucro líquido |
|---|---|---|
| Ano 1 | R$ 0 | R$ 0 |
| Ano 2 | R$ 0 | R$ 0 |
| Ano 3 | R$ 0 | R$ 0 |
Preencha o negócio e os números, depois escreva o plano.
Procure um modelo de plano de negócios e você recebe um documento em branco. Isso servia quando o difícil era escrever o texto. Não é mais.
O difícil são os números. Um banco vai primeiro à projeção, e o que ele confere leva dois minutos: cada venda dá lucro? quantas vendas por mês cobrem os custos fixos? qual é o ponto mais baixo do caixa? e o valor pedido cobre esse buraco? Um plano que falha nessas quatro perguntas não é lido além delas.
Aqui a projeção é aritmética sobre os seus próprios números, atualizada enquanto você digita, e as seções escritas recebem esses números como fatos que devem usar e não podem inventar.
Say who will read it
A bank, an investor, a visa caseworker or yourself. A lender wants to know the loan can be serviced in a bad month; an investor wants to know how big this gets. The same business needs two different documents.
Describe the business
What it is, who buys it, what problem it solves, who else does it. Plain answers; the writing comes later.
Enter the unit economics
Price per sale, direct cost per sale, opening volume, growth, fixed costs. The projection, break even volume and cash trough appear immediately, before anything is written.
Read the check, then write it
The check names what a reader will notice: a negative unit margin, a month where the cash runs out, a growth rate that compounds into a number nobody believes. Fix those, then generate the plan.
They take about two minutes, and a plan that fails them is not read past them.
Where a template leaves it
What this does
Every finding is arithmetic on the numbers you entered, so it can be argued with rather than believed.
A negative unit margin
If each sale loses money, no amount of volume fixes it, and nothing else in the plan matters. This is checked before anything else.
The month the cash runs out
Annual figures hide it. A business with a profitable year one can still stop in month four, and that is the month a lender is looking for.
A funding gap
Startup cost plus the cash trough is what the plan actually needs. If the ask is smaller than that, the plan is asking for the wrong number.
Growth that compounds too hard
Six per cent a month sounds modest and is eight times bigger by year three. The compounded figure is what gets tested, not the monthly rate.
A thin gross margin
Below twenty per cent there is very little room for a bad month, a price rise from a supplier, or a customer paying late.
No fixed costs and no competitors
Both read as not having looked. Rent, software, insurance and your own time all belong in fixed costs, and "no direct competitors" is the sentence lenders least believe.
Why the model never touches the numbers
A language model will happily produce a projection. It will be internally plausible, formatted well, and wrong in ways that are hard to see, because it is generating text that looks like a spreadsheet rather than doing arithmetic. That matters more here than almost anywhere else, because the projection is the part a reader checks. An investor who finds one figure that does not reconcile with another stops trusting the whole document, and they are right to. So the split is strict. Every number on this page and in the exported PDF comes from arithmetic on what you entered: margin, break even volume, monthly cash, the three year summary, the funding gap. The written sections are then given those figures as facts, with an instruction never to invent one and to refer to the real ones where they matter. What you get is a plan whose prose and whose tables agree, which is not the normal outcome of asking a chatbot for a business plan.
Assine para gerações ilimitadas ou compre créditos, pague só pelo que usar. Cancele quando quiser.
