The early-stage forecasting challenge
Established businesses forecast revenue by extrapolating trends from years of historical data. Startups and early-stage companies do not have that luxury. With six months of revenue history -- or none at all -- traditional statistical methods are unreliable.
Yet early-stage companies need revenue forecasts more than anyone. Investors demand them. Hiring plans depend on them. Cash runway calculations are meaningless without them.
Bottom-up beats top-down
When history is limited, the top-down approach ("the market is £5 billion, we will capture 1%") produces numbers that sound impressive but mean nothing. Investors see through this immediately.
Instead, build bottom-up from the activities that generate revenue:
For B2B SaaS: How many qualified leads per month? What is the conversion rate from demo to close? What is the average contract value? How long is the sales cycle? Multiply these together and you have a revenue forecast grounded in operational reality.
For transactional businesses: How many customers per day/week/month? What is the average transaction value? What is the repeat purchase rate? These are observable even with limited history.
For marketplaces: What is the gross merchandise value? What is the take rate? How fast is seller/buyer acquisition growing?
Using analogues and cohorts
When your own data is thin, borrow context from comparable companies:
Public company filings. SaaS companies that IPO'd publish their early growth trajectories in S-1 filings. While your path will differ, these provide useful reference points for growth rates at similar stages.
Industry benchmarks. Organisations like SaaS Capital, KeyBanc, and OpenView publish median growth rates by company stage and size. These help you sanity-check your assumptions.
Cohort analysis. Even three months of data can reveal patterns when you analyse by cohort. Do customers who sign in January behave differently from those who sign in March? Is the second-month retention rate stable? Cohort patterns stabilise faster than aggregate trends.
Building credibility
A revenue forecast is only useful if stakeholders trust it. Three practices build credibility:
Show your assumptions explicitly. Present the drivers alongside the output. When a board member questions the revenue number, you can point to the specific assumption they disagree with and model the alternative on the spot.
Provide a range, not a point estimate. Early-stage forecasts should present a base case with clearly defined upside and downside scenarios. This demonstrates intellectual honesty and gives decision-makers a framework for risk assessment.
Update monthly and track accuracy. Compare each month's forecast to actuals, and adjust the model. When stakeholders see the forecast improving over time, trust builds naturally. A forecast that was 30% off in month one but 10% off by month six tells a story of learning, not incompetence.
Common mistakes
The most dangerous mistake is confusing a revenue target with a revenue forecast. A target is what you want to happen. A forecast is what you expect to happen given current assumptions. When the two diverge, the forecast should update -- the target should be revisited separately.