You start a business with no money by selling a service or a pre-order first, something that converts your time into revenue before anything needs buying. Product businesses that need stock, tooling or a licence are the wrong shape for a zero budget, no matter how good the idea is. Here is the order, and what each stage costs.
What "no money" actually means
There is no zero-cost business. There is only a business whose cost at this stage is measured in hours instead of dollars, and being honest about that changes the choice of business more than any idea does.
Put two shapes side by side. One business needs 400 hours before its first sale. Another needs $200 and a fortnight. For most people the second is the cheaper business, including people who do not have the $200. Hours are what you are actually spending, and you do not get this year's hours back.
That makes the first question arithmetic: how many hours between today and the first payment? Ask it of every candidate. A shape that answers "a weekend" beats one that answers "three months," whatever either one costs in money, because the hours are the real budget and they are already being spent.
Two follow-up questions belong in the same set. What must exist before someone can pay? And what would make you stop? Write all three answers down before choosing. On paper, almost every idea looks cheaper than it is.
If you do not have a candidate yet, the business idea generator interviews you about the skills, hours and constraints you actually have, and comes back with ideas shaped around them. Take its output through the three questions above before you spend an hour on any of them.
What to sell first: a service, a pre-order, or a promise
Three shapes work at zero budget, and they share one property: money moves before costs do. Every entry on the 12 unique business ideas list can be tested in one of these shapes, and so can most ideas you will ever have.
- A service. You sell a result and deliver it with hours you already own: a fixed-scope audit, a done-for-you version of a job you have been doing badly for yourself, a recurring chore someone is glad to hand over. The first sale is an invoice for work you delivered by hand to someone who was free to say no.
- A pre-order. You sell the thing before it exists: a product at a price with a delivery date attached, or a deposit that holds a place in the first run. The first sale is money taken against a promise, and the delivery is where you find out what the business actually costs to run.
- A promise. You sell a commitment rather than a deliverable: a paid workshop before a course exists, a maintenance contract for machines you know inside out, a booked date held with a deposit. The first sale is a filled place in a calendar, which is the evidence that the rest of the calendar can fill.
The fourth shape does not work at zero budget: anything you must buy or build before someone can pay. Stock, tooling, a licence, a finished product. Each of those puts a cost in front of the first payment, and no quality of idea changes that order.
If that is your idea, none of this is permission to skip ahead and start building. Find the smallest slice of it somebody can pay for before the expensive part exists: the pre-order, the deposit, the hand-delivered version, the consulting shape of the product. Whether anyone wants this still gets answered first. Only the order of the spending changes.
What all three first sales have in common is that a stranger's money settled something that conversation could not. Price the outcome rather than the hours, deliver the first one by hand and slowly, and let the second customer teach you what to fix. As an example: 30 conversations → 5 quotes → 2 sales at $200, so 50 hours of outreach and delivery returns $400. Replace these with your counts. The figure you want is how many conversations turn into payments, and at what price.
The order to spend your first $100 in
If some money exists, spend it in the order of what it buys you: information first, tools last. Most early spending goes to things that make the founder feel like a business rather than things that answer a question.
Where the first $100 goes
A domain and one page
A real address, and one page carrying a single promise, a price and one action. This is what every later test measures.
The cheapest test of the riskiest assumption
Ask the question that would sink the idea first. Conversations cost hours. A small run of paid traffic costs dollars and returns a clean read on demand on your own schedule.
Tools, once a task repeats
Buy a tool when doing the job by hand costs more hours each week than the tool costs each month. Not before.
If the budget is genuinely zero, the order does not change and neither does the work. Only the unit changes: you spend the hours directly.
The manual path: your first two weeks
What is genuinely free, and what the catch is
Free tiers, free trials and free tools all charge something. The bill arrives as your data, your switching cost, or your time later. Name the trade before you take the free thing, and most of these stay good deals.
| What is free | What it charges |
|---|---|
| A free tier of a paid tool | Your data, then a price when moving costs more than paying |
| A free trial | Setup hours you do not get back |
| A site on a free subdomain | Credibility with buyers, and a move later |
| Advice from people who like you | Answers that cannot come back as a no |
| Posting on social platforms | Reach the platform can change or remove |
The data trade is the quiet one. The free tier that holds your customer list also holds it at the moment you want to leave, which is usually the moment the business starts working. The switching cost is the same trade seen from the other side: the longer you stay, the more of your work lives inside someone else's product.
The time trade is the loud one. A free tool with no budget behind it gets replaced, rebuilt and relearned, and the relearning is unpaid. Keep a rough count of the hours. When a month of them exceeds what the paid version costs, the free version is the expensive one.
A domain is the cheapest thing here to buy and it removes the biggest catch. The address stays yours, so the page you build this month is the page you keep, whatever tool you build it with.
When the free path stops working
Free methods cost hours and paid methods cost dollars, so the switch point is a comparison rather than a milestone. When the hours a free route takes are worth more than the money the paid route costs, spending is the cheaper option.
One question locates the point: how many hours does it take you to put one promise in front of a hundred of the right strangers for free? If the honest answer is twenty hours of posting and replying, and a small ad run reaches the same hundred in an afternoon for a few hundred dollars, then anyone whose hour is worth more than the difference is paying extra to stay free.
There is a second marker: reach that runs out. Posting into communities works until the community has seen you twice. Past that point the channel is not slower, it is finished, and paid traffic is the channel still accepting new customers this week. It gives a clean read on demand on your own schedule, and for a market with nowhere public to gather it is the primary channel rather than a fallback.
This is where the free validation methods hand over to paid testing. If the budget is genuinely zero, those free methods are the right start. The paid test is the fastest version of the same question: a few hundred dollars of traffic puts one promise in front of strangers this week and tells you whether they act on it.
What not to do instead of having money
Before the first sale, borrowed money is a bet on a question you have not asked yet.
- Credit cards. The most available option and the most expensive one. A card turns a wrong guess into a balance that grows while you learn.
- Loans. Lenders ask for the plan that testing was supposed to produce. A bank approving a forecast is not a customer paying a price, and only one of those two proves anything.
- Grants. The most-searched option and the worst first move. Grant programmes are slow, competitive and written for businesses that already know what they are selling. A panel funds a plan, not a question. Run your idea through the market size calculator first, so you know what a grant would be funding and whether the ceiling is worth the weeks an application costs.
Borrowing does not answer the question of whether anyone will pay. It only changes who pays for the asking. After a first sale the same instruments change meaning: a loan against real orders is ordinary finance, and the same loan before any orders is debt on a guess.
What to do from here
Two things are worth doing in whatever shape you pick. Work through the starting a business checklist in the order it sets out: proving demand, pricing and one sale come before registration, and that order is the cheapest way to find out you were wrong. Then run the loop by hand a few times before you pay anything to run it for you.
When you want the loop run with you rather than alone, Draper does that step by step. The first two workshops are free and need no card.
