How to Forecast SaaS Revenue (Without Guessing)

Ask a SaaS founder for their revenue forecast and you'll often get one number — a straight line up and to the right. Ask a SaaS investor what they think of that forecast and you'll usually get a raised eyebrow, because a single-point forecast is a guess wearing a spreadsheet's clothing. A real forecast is a range, built from a small number of assumptions that actually drive the business.

The four inputs that actually matter

Underneath every SaaS revenue forecast, however complex the spreadsheet looks, there are really only a handful of drivers worth getting right:

Starting MRR (Monthly Recurring Revenue). Your current baseline — the one input that isn't a guess.

New customer growth rate. How fast you're adding new paying customers or new MRR each month, usually expressed as a percentage growth rate or a flat number of new customers.

Churn rate. The percentage of MRR (or customers) you lose each month. This one has an outsized effect on long-run forecasts because it compounds against you the same way growth compounds for you.

Expansion revenue. Revenue growth from existing customers upgrading, adding seats, or moving to higher tiers — separate from new customer acquisition, and often underestimated in early forecasts.

Why a single-point forecast is close to useless

A forecast built on "we'll grow 10% month over month" is really a forecast built on one specific, optimistic assumption holding true for every single month of the forecast period — which almost never happens in practice. Growth rates fluctuate, churn spikes in bad quarters, and expansion revenue is lumpy. A single-point forecast doesn't communicate any of that uncertainty; it just presents one path as if it were the only plausible one.

Bear, base, and bull: building a real range

The standard fix is to forecast three scenarios instead of one, using the same model with different assumptions plugged in for growth and churn:

Bear case — conservative growth, higher-than-current churn. This is your "if things go worse than expected" floor, useful for runway planning and worst-case cash decisions.

Base case — growth and churn roughly in line with your recent actual trend, extrapolated forward. This is the number you'd actually plan around operationally.

Bull case — improved growth, reduced churn. Useful for understanding upside, but not what you should be budgeting against.

The gap between the bear and bull case tells you how sensitive your business actually is to small changes in growth and churn assumptions — a wide gap means the business is high-variance, and a narrow gap means the forecast is more robust to being slightly wrong.

A simplified worked example

Start at $50,000 MRR. In the base case, assume 8% month-over-month new growth and 3% monthly churn — net growth of roughly 5%/month. Over 12 months compounding monthly, that reaches roughly $89,000 MRR. In the bear case, drop new growth to 5% and raise churn to 4%, for roughly 1%/month net growth, reaching only about $56,000 MRR over the same period — a dramatically different outcome from what looks like a small assumption change. That sensitivity is exactly why the three-scenario approach matters more than getting a single number "right."

What forecasts commonly get wrong

The most frequent mistake is applying a flat blended churn rate across the whole customer base, when in reality churn is usually concentrated in a specific cohort (often newer or lower-tier customers) and much lower elsewhere — a blended rate can understate risk in the near term and overstate it once the business matures. The second most common mistake is forgetting expansion revenue entirely, which for a healthy SaaS business is often responsible for a meaningful share of net new MRR growth, not just new customer acquisition.

To run your own bear, base, and bull projections — including ARR and LTV — the free Revenue Forecast Calculator builds all three scenarios from the same growth and churn inputs described here.

Common SaaS forecasting mistakes

A frequent mistake is forecasting new customer growth using the same churn rate that applied to an earlier, smaller customer base — churn often increases as a company scales into less ideal-fit customer segments, and using an outdated, lower churn assumption from an earlier stage can significantly overstate the forecast.

Another common error is ignoring the difference between logo churn (customers lost) and revenue churn (dollars lost) — a company can have low logo churn but high revenue churn if it's disproportionately losing its largest accounts, a distinction that matters enormously for an accurate forecast but is easy to miss if only one churn metric is tracked.

A simplified worked example

A SaaS company starts the year with 200 customers paying $500/month average ($100,000 MRR). Assuming 15 new customers per month, 3% monthly churn, and no expansion revenue: by month 12, the base grows to roughly 260 customers and $130,000 MRR in the base case.

Running the same model with churn at 5% instead of 3% — a plausible bear-case shift — the ending customer count drops to roughly 220 and MRR to about $110,000, a meaningfully different outcome from a churn assumption that might look like a minor detail in a spreadsheet. This is exactly why the article emphasizes churn assumptions as the most consequential single input in a SaaS forecast.

A practical forecasting approach

Rebuild your churn assumption using your most recent 3-6 months of actual data, not an assumption inherited from an earlier, smaller stage of the business — churn tends to drift as a company scales, and stale assumptions are one of the most common sources of forecast error.

Track logo churn and revenue churn as two separate numbers, not one combined figure — a forecast that only tracks customer count can miss a dangerous trend of disproportionately losing your largest accounts even while overall customer numbers look stable.

Why cohort-based forecasting outperforms simple growth extrapolation

Rather than simply extrapolating an overall growth rate forward, tracking revenue by customer cohort (grouping customers by when they signed up, and watching how each cohort's revenue evolves over time) tends to produce more accurate forecasts, since it captures how churn and expansion actually behave differently for newer versus more established customers.

This more granular approach takes more setup than a simple top-line growth extrapolation, but becomes increasingly valuable as a company scales and cohort behavior starts to meaningfully diverge from a single blended average.