SaaS revenue forecasting, done properly.
How to forecast next month's MRR from your Stripe data, and why a range beats a single number.
The simple version.
Next month's revenue is this month's recurring revenue, minus what cancels, plus what's new. Everything else is about estimating those two moving parts well: how much will cancel, and how much will arrive.
next month ≈ MRR − (MRR × churn rate) + new and expansion revenue
Three ways to forecast, and what each misses.
Last month again
Next month = this month.
Misses: Any growth or churn. Fine for a very stable business, wrong for one that's changing.
Straight-line trend
Fit a line through the last 6 to 12 months and extend it.
Misses: Turning points. A launch spike or a churn wobble drags the line in the wrong direction for months.
Customer by customer
Work out who is likely to renew, cancel or upgrade, and how many new customers usually arrive, then add it up.
Misses: Very little history. With only a few months of data, any method has a wide range, and an honest one shows it.
Test it before you trust it.
The fair test of any forecast is to hide the last few months, forecast them from what came before, and compare. If a method can't predict months you already know, it won't predict the next one. Use the typical miss on those tests to set your range.
Where Scenario fits.
Scenario connects to Stripe read-only, forecasts customer by customer, tests itself on months it wasn't trained on, and shows next month with a likely range and what's driving it. A forecasting engine does the maths; the AI only explains it. Customer names never go to the AI.
How do I forecast next month's SaaS revenue?
Start from this month's recurring revenue, take away what's likely to cancel, and add what's likely to arrive: new customers and upgrades. £10,000 of MRR with 4% monthly churn and £800 of new revenue a month gives about £10,000 − £400 + £800 = £10,400 next month. A better forecast works customer by customer and gives a range, not one number.
What's the most accurate way to forecast MRR?
Forecasting from your own customers' behaviour (who renews, who cancels, how many new customers arrive, and when) is usually more accurate than drawing a line through past totals, because it notices changes underneath the total. Whatever the method, test it on past months it didn't see, and judge it by how close it got.
Can I forecast revenue from Stripe?
Yes. Stripe holds the subscriptions, renewals, cancellations and payments a forecast needs. You can export it to a spreadsheet, or connect a tool like Scenario, which reads Stripe read-only and keeps the forecast up to date.
Why does my forecast need a range?
Because next month isn't certain: a few customers more or less can move it. A range (for example, likely between £9,900 and £10,800) tells you how much to rely on the middle figure, and whether a bad month is still survivable.
See a real forecast in a minute.
The live demo runs on a sample business, with no sign-up. Or connect Stripe and see your own next month.
Related: Runway calculator · Churn calculator · Price increase calculator