How to forecast revenue with little or no sales history

With no closed deals behind you, every method that derives probabilities from historical conversion rates is unavailable. The answer is not to borrow someone else's benchmarks. It is to forecast the handful of deals you do have one at a time, report a range rather than a number, and start recording what you called so that in two quarters you have the history you are missing.

Why industry benchmarks will not save you

The tempting move is to find a published win rate for your category and apply it. It fails for a specific reason: those figures are averages over companies whose sales motion, price point, buyer and stage definitions have nothing to do with yours. A 22% average win rate across SaaS tells you about SaaS, not about whether this buyer signs in March.

Worse, it feels like rigour. A forecast built on a borrowed conversion rate carries a false authority that your own honest guess does not, and false authority is precisely what you cannot afford when the error bars are this wide.

Forecast each deal, not the funnel

With six or ten open deals you are not in statistics territory, which is liberating. You can hold every deal in your head. So reason about each one directly: what has the buyer actually done that cost them something, do you know who signs, and what would tell you it is dead.

Then put a number on it. Your first numbers will be badly calibrated, and that is fine — they are the raw material for the calibration you do not have yet.

The honest answer here is a span of 145.000 € on a weighted figure of 92.000 €. That is not a failure of method. That is what six deals of uneven size genuinely imply, and a board that hears it will make better decisions than one told a single confident number.

Early-stage forecasting is not less accurate because you lack tooling. It is less accurate because six coin flips have a wide distribution. The job is to report that honestly, not to hide it behind a point estimate.

Give two numbers, not one

Investors and boards handle ranges better than founders expect, provided you say what drives them. "We land between 35.000 € and 180.000 €, and it mostly depends on the 120.000 € enterprise deal" is a sentence that invites a useful conversation. A single number invites a false one.

If you must give one figure, give the conservative end and treat anything above it as upside. Being right about a low number twice buys you more credibility than being nearly right about an optimistic one.

Start the record now, because it is cheap and it compounds

• One row per open deal per month: date, deal, value, the odds you gave it, the month you expected it to close.
• When it resolves, add what happened and when. Do not overwrite the original call — that call is the data.
• After two quarters, compare. Of the deals you called around 60%, how many closed? That number is your first real conversion rate, and it is about your business rather than a category average.

This is the whole trick, and it costs a spreadsheet. A year from now the question "what is our win rate at this stage" has an answer derived from your own deals, and every forecast after that is better than the ones you are making today.