Every Budget System Assumes a Salary. Here Is What I Do Instead.
Twelve years of freelance income means some months are triple others. The method that finally worked treats the good months as a reservoir rather than as income.
Almost every budgeting method starts from a monthly figure that arrives on a predictable date. Mine does not. Over the last three years my best month was 4.1 times my worst, and the worst month was not the one with no work — it was the one where three clients happened to pay late at the same time. Any system built on “your monthly income” fails on contact with that, and I spent years assuming the failure was mine.

What works for me is to stop treating incoming money as income at all. Everything that arrives goes into one account that I think of as a reservoir, and once a month I pay myself a fixed amount from it into the account I actually spend from. That fixed amount is my income, as far as every budgeting decision is concerned, and it does not change when a good month happens.
Setting the number is the whole difficulty and I got it wrong twice. My first attempt was my average monthly income over the previous year, which failed within four months because an average is exceeded roughly half the time and the reservoir drained during a bad run. My second attempt was the minimum, which was technically safe and so low that I ignored it. What has worked for three years is roughly the thirtieth percentile of my monthly income over the trailing two years — a number I exceed about seven months in ten.
The reservoir needs a floor before this works at all, and I would be honest that this is the hard part and the reason the method is not available to everyone immediately. Mine is three months of the fixed payment. Getting there took about fourteen months of paying myself slightly less than I could have, and during that period the method felt like a worse version of just spending what arrived. It only starts working once the buffer exists, which means the first year is the tax you pay for every year after.
What it buys, concretely: I have not thought about cash flow since 2023. A client paying sixty days late is now a reservoir event rather than a personal one. I no longer take work I do not want because of when the money lands, which has probably had a larger effect on my income than any budgeting decision.
Two adjustments I make. Once a year, in January, I look at the reservoir and if it is more than about five months deep I raise the fixed payment. This is deliberately slow — raising it is easy and lowering it is miserable, so I want a full year of evidence before I do. And I keep tax entirely outside the reservoir, in a separate account, moved on the day money arrives rather than at any point later. That is not sophistication; it is that I once did it the other way and it was the worst financial week of my life.
The thing that surprised me is psychological rather than mechanical. When a good month happens now, I do not feel richer, because nothing changes in my spending account. What I feel instead is the buffer getting deeper, which is a quieter feeling and a much more useful one. Under my old system, a strong month produced a spending month roughly three-quarters of the time — I checked, going back through my records — and I would have denied that if you had asked me.
I am not an adviser and my situation is specific: no dependants, and a line of work where the worst case is a lean quarter rather than nothing. What I would say generalises is only the framing — if your income is irregular, the useful question is not how to budget it but how to convert it into something regular first.


