TL;DR
Yes, you need one sentence that says who pays, for what and how often, before you launch. Without it, you cannot set a price, size a market or count the customers you need, so every number after that is a guess. Subscriptions, one-off sales and solo services are the easiest models to check on paper.
Yes. You do not need a business plan, but you need one sentence that says who pays you, for what, and how often. Without that sentence you cannot set a price, size a market or count the customers you need, so every number you produce afterwards is a guess.
What a revenue model means here
Forget the canvas with nine boxes. For a solo founder, a revenue model is one line you can fill in:
Someone pays an amount for something, this often.
Three shapes fill that line cleanly:
- Subscription. A customer pays €19 every month for access to the product.
- One-off sale. A customer pays €49 once for a template, a tool or a report.
- Solo service. A client pays €600 for a project you deliver yourself.
Other shapes exist: advertising, commissions on transactions, billing by usage. They are real business models, but they do not fit on that line as neatly, because the amount depends on volumes you do not control yet. More on that below.
If you cannot fill the line at all, that is the answer to the question in the title.
"No model yet" is more common than you think
At Banast, we have analysed 225 projects so far. This is our own count, not a study, but the split is telling: 39% were subscriptions, 9% one-off sales, 9% services, and 22% had no revenue model at all. The remaining 21% is everything else.
So more than one project in five had a working product and no stated way of earning from it.
That is not a harmless gap. In CB Insights' review of 431 funded startups that shut down since 2023, 43% cited poor product-market fit (CB Insights, 2026). Not knowing who will pay you, and for what, is one of the most direct ways to end up there.
A worked example: one target, three models
Say you want your project to bring in €3,000 a month in revenue. Here is what that target means under each of the three simple models.
Subscription at €19 a month. 3,000 ÷ 19 = 157.9, so 158 paying subscribers, all at the same time. And every one who cancels has to be replaced just to stay there.
One-off sale at €49. 3,000 ÷ 49 = 61.2, so 62 new sales. Not once: every single month, because last month's buyers do not pay again.
Solo service at €600 a project. 3,000 ÷ 600 = 5 projects a month. If each one takes you three days, that is 5 × 3 = 15 working days a month, before any time spent finding the next client.
Same target, three very different businesses. The first needs you to keep customers. The second needs a steady flow of new buyers. The third needs your calendar more than your marketing. Each one tells you what to work on next.
Now try it with no model. €3,000 divided by what? Users? Sign-ups? Visits? There is no number to divide by, so there is no target, and no way to tell whether you are getting closer to it.
Take advertising as the contrast. Monthly revenue there is roughly:
visits × ads seen per visit × what advertisers pay per 1,000 ads ÷ 1,000
That is three numbers instead of one, and before launch you know none of them. The model is not wrong. It is just not checkable on paper yet, which makes any projection built on it fragile.
These figures are revenue, not profit. Payment fees and running costs come off the top, which is why the break-even math for quitting your job starts from margin instead. You can run your own numbers in the break-even calculator.
Why "I'll figure it out later" costs more than it looks
The revenue model is not one decision among others. Most of the decisions after it depend on it.
- Price. You cannot pick €19 or €49 until you know whether it is per month or per purchase.
- Customer target. As the example shows, 158 subscribers and 62 sales a month are different jobs.
- Channels. A one-off product needs new traffic every month. A subscription can afford to spend more to win each customer, because that customer keeps paying.
Leaving the model for later does not keep your options open. It means the price, the target and the channels you choose in the meantime are not anchored to anything, and you will redo them.
How to pick one this week
- Write the sentence. Fill in the line above for your project, even if you are unsure. A wrong sentence can be corrected. A blank one cannot.
- Run the division. Pick a monthly target and divide it by your price, as in the example. Look at the result and ask whether it is a number you can reach.
- Check what it asks of you. Retention, a flow of new buyers, or your own time. If you hate the answer, try another model on paper before you write any more code.
- Put a price on the page. A visible price is the cheapest test of the sentence. If you have no data to set it, start with how to price a micro SaaS with zero customers.
What this method does not cover
Choosing a model does not prove anyone will pay. It tells you what to measure, not what the result will be. Only real buyers answer that.
It also fits simple projects best. Marketplaces, ad-funded products and usage-based billing need more inputs (traffic, take rate, consumption per customer), and a single division will mislead you there.
And the example ignores fees, costs, taxes and churn. Treat the counts above as a floor, never as a plan.
How Banast does it
Before paying for a report, every founder ticks how their project earns money. Three cases are accepted: subscription, one-off sales and solo services. Advertising, commissions, usage-based billing, projects with no revenue model yet, and companies of more than five people are turned down. The free pre-audit stays available to everyone either way.
The rule uses no AI. It is about ten lines of code, and it came out of those 225 projects. On the three accepted models, our calculations hold. On the others, they would be approximate at best. We would rather lose a sale than sell a report we know could be wrong.
FAQ
Can I launch for free and pick a model later?
You can launch for free, but pick the model now and write it down. Free users tell you the product works, not that anyone will pay, and the model decides which numbers you should be watching while they use it.
My project makes money from ads or commissions. Does this still apply?
Yes, but the math needs more inputs. Ads depend on traffic and the price advertisers pay, commissions on the volume other people trade through you. You need an estimate for each before the model says anything useful.
Can I change my revenue model after launch?
Yes, and many projects do. Changing a model you wrote down is a decision. Discovering you never had one is a surprise, and usually a late one.
Adrien Lebas
Founder of Banast, building it alone from Avignon. Writes about what it takes to make a living from what you build.