This is the worked version. It assumes you already accept that averaging a lumpy year hides the only thing worth seeing, and want to actually build the thing on your own numbers this afternoon. The reasoning behind it, and why the four layers must never be added together, is a separate piece: how to forecast revenue when you do not bill the same amount every month.
Allow an hour the first time. Fifteen minutes a month after that.
What you need in front of you before you start
Four numbers and one list. Every one of them already exists somewhere; the work is putting them side by side rather than calculating anything clever.
| What | Where it lives | What it is not |
|---|---|---|
| Trailing six months billed | Your profit and loss, month by month | Not what you collected. What you invoiced |
| Outstanding invoices | Aged receivables | Not the overdue ones only. All of them |
| Contracted, not yet billed | Your signed agreements, instalment by instalment | Not the whole contract value. Only the rows you have not invoiced |
| Open proposals | Your pipeline, one row each | Not a total. One line per proposal, with its own odds |
In RunwayDue the first two come off reports, the third off the schedule on each contract, and the fourth off proposals. Anywhere else, a spreadsheet with four columns does the same job. The model does not care where the numbers come from, only that they stay apart.
Step one: the trailing average, which is the only average you are allowed
Add up what you invoiced in each of the last six months and divide by six.
This is the one place averaging is legitimate, and it is worth being clear about why. You are not using it to predict anything. You are using it as a unit of measurement: a rough answer to "what does a normal month cost me to survive". Every figure after this is expressed in those units.
Six months rather than twelve because a year of a growing business averages in a version of it that no longer exists, and three months is too short to survive one good invoice.
Step two: weight the pipeline once, before it touches anything else
Take each open proposal and multiply it by your honest chance of winning it. Not the client's enthusiasm. Your history with work that looked like this.
The discipline is doing it per proposal and never afterwards. A single "pipeline: $98,000" figure that you privately discount by some feeling is a number you will act on at full value the first week you are worried about cash.
Step three: secured months, which is the number you actually watch
Add outstanding invoices to contracted-not-yet-billed. Divide by your trailing average.
That is how many months of normal trading are already committed to you, ignoring everything you have not won. It is one number, it moves slowly, and it tells you when to sell rather than whether last month was good.
Pipeline stays out of it deliberately. The moment a maybe enters this figure it stops being the thing you can rely on and becomes the thing you hope for, and you will not be able to tell which you are looking at three months from now.
A worked example, end to end
Six months of billing: $18,000, $4,000, $31,000, nothing at all, $12,000, $25,000. That totals $90,000, so the trailing average is $15,000 a month. Note the empty month sitting in there. It belongs in the average; it is part of the shape.
Outstanding invoices come to $22,500. Contracted and not yet billed, adding up the instalment rows nobody has invoiced: $36,000.
Three proposals are open. $30,000 at sixty percent is $18,000. $18,000 at twenty-five percent is $4,500. $50,000 at ten percent is $5,000. Weighted pipeline is $27,500, against a raw total of $98,000. That gap is the entire reason to do this.
Secured months: $22,500 plus $36,000 is $58,500, divided by $15,000, is 3.9 months.
What each answer actually changes
The point of keeping them apart is that each one answers a different question, and you will have all four questions in a normal quarter.
- Can I pay everyone this month? Collected, plus the invoices falling due inside thirty days. Nothing else is money.
- Can I take on that project? Contracted. If it is already thin, a new project is capacity you have promised twice.
- Do I need to be selling right now? Secured months. Under three is a prompt, whatever the pipeline looks like.
- Is the pipeline real? Compare weighted pipeline against what you actually won last quarter. If you have been weighting at fifty percent and closing at twenty, the model is fine and your odds are wrong.
Doing it again next month
Same day each month, and write the number down. The absolute figure matters much less than the direction.
Three things are worth watching across a few months. Secured months trending down while pipeline grows means you are selling but not closing. Secured months flat while the trailing average climbs means you are growing into your own commitments and will need to sell harder than it feels like. And a month where outstanding invoices jump without new work means you are not being paid, which is a collections problem wearing a forecasting costume.