Predicting your next 90 days of renewals without wishful thinking

The decision is small and it will not leave you alone. Hire an editor for four hours a week. Pay for the better camera. Book the flights in October. Each one needs the same input — what will be arriving each month between now and then — and you do not have it, because the only number you have is last month's total, which included a launch and a refund and one very good Tuesday. So you postpone the decision, which is itself a decision, and usually the wrong one.

Why last month's total tells you almost nothing

A monthly income figure is a mixture of things with completely different half-lives. A course sale is over. A tip is over. A renewal, on the other hand, is a statement about next month — the same person is likely to pay again, and the month after that, until something changes. Adding them together produces a number that cannot be extended forwards, which is why forecasting from a total always feels like guessing. It is guessing.

The genuinely difficult part is that early recurring data is noisy. With 40 members, two people leaving in one month is a 5 per cent rate and three leaving is 7.5 per cent, and neither figure means anything on its own. Small numbers wobble. You need three or four months before your cancellation rate is describing your business rather than describing coincidence, and there is no honest shortcut.

The second difficulty is emotional. Nobody enjoys writing down a number that goes downwards, and a base forecast always goes downwards, because it assumes not one new member joins. That is the point of it. The base is the floor you stand on if every piece of marketing you plan this quarter fails completely, and knowing that floor is what allows you to make commitments in the first place.

Two ways to answer "what will I earn in October?"

One method uses the whole of last month and a feeling. The other uses only the part of last month that will happen again.

Building the forecast in about half an hour

Open your per-product analytics and a blank page. You are producing three numbers per recurring product, then one line per month for the quarter ahead.

1. 1Count active members for each recurring product today. Not total sign-ups since launch — the ones currently paying.
2. 2Find your monthly cancellation rate: members who left last month, divided by members at the start of last month. Do this for three separate months and take the middle value rather than the kindest one.
3. 3Convert yearly members into a monthly figure by dividing their annual price by twelve, and note the month each renewal falls due — those are single points of risk, not a smooth line.
4. 4Multiply active members by price to get this month's base. Then apply the cancellation rate to get next month's base, and again for the two months after that.
5. 5Write the three resulting monthly figures down. That is your quarter, assuming you never sell another subscription.
6. 6Now add a growth line, separately and conservatively: the number of new members you have genuinely averaged over the past three months, not the number from your best month.
7. 7Plan every fixed commitment — software, help, rent, the flights — against the base line only. Use the growth line for things you can cancel.
8. 8Repeat on the first of each month and compare last month's prediction to what actually happened. Two cycles of that and your rate stops being a guess.

A quarter, worked through

Take a membership at $12 a month with 150 active members, and suppose you have measured your own cancellation rate at 6 per cent a month across three months. Ignore new joiners entirely. The base looks like this.

So the floor for the quarter is roughly $4,788 — that is the money that arrives if every launch fails and no new member ever appears. If you have also been adding an average of 14 members a month, the realistic line sits meaningfully above the floor, but you commit at $1,500 a month and treat the rest as upside. This is a worked illustration rather than a forecast for your business; your rate and your prices will produce different figures. What transfers is the discipline: the floor pays for the commitments, and the growth pays for the ambitions.

The mistake most people make

They forecast growth and forget decay. The sheet says 150 members plus 15 a month equals 195 by the quarter's end, which would be true only if nobody ever left. Every recurring business loses members constantly; it is not a failure state, it is the weather. Build the losing side into the model first, and treat every joiner as something that offsets it. The other version of the mistake is checking the forecast only when things feel wrong. A rate that drifts from 5 per cent to 9 per cent over two months is the single most useful early signal you will ever get, and it is invisible unless you look monthly. If cancellations are climbing, the answer is usually in what people were promised at sign-up rather than in the price — worth reading the FAQ on how access and billing work before you start changing numbers.

A forecast is not a promise to yourself about how well things will go. It is the smallest number you can safely build on, written down where you can see it, updated once a month. Do that for two quarters and the decisions that currently sit unresolved in the back of your mind — the editor, the camera, the flights — start answering themselves. Your customer list exports to CSV on every plan, including Free, if you would rather keep the whole sum in a spreadsheet of your own.