Why 260 working days was never the number
Almost everyone sets a first freelance rate the same way: take the old salary, halve it or not, divide by 220, call it a day rate. Two years later they are working more and keeping less than when they were employed. The division was never the problem. The 220 was.
Where the days actually go
Work it down one line at a time, with an invented but complete set of inputs.
- A full year of a five-day week:
52 × 5 = 260days. - Take five weeks of holiday:
52 − 5 = 47weeks, so47 × 5 = 235days. - Take 8 public holidays and 5 days set aside for being ill:
235 − 8 − 5 = 222days you can actually work. - Now the one that decides everything. Quoting, invoicing, chasing payment, marketing, bookkeeping, admin and the calls that go nowhere are all real work and none of it is chargeable. At a 75% billable share:
222 × 0.75 = 166.5days.
So 260 becomes 166.5, and 220 was never within reach. Track your own weeks for a month or two before you pick that percentage — it is your number, not a norm, and this page has no opinion about what it should be.
Then work backwards from what you need
Say the target take-home is $60,000, with $6,000 a year going into a pension nobody is paying for you any more, and a tax-and-contributions rate you looked up yourself of 30%.
You need 60,000 + 6,000 = $66,000 after tax. Grossing that up: 66,000 ÷ 0.70 = $94,285.71 of profit before tax. Add business costs — software, insurance, equipment, an accountant, workspace, training, travel — at $9,000: the year has to take in 94,285.71 + 9,000 = $103,285.71.
Spread over the billable days: 103,285.71 ÷ 166.5 = $620.34 a day, rounded up to the cent, because a rate rounded down lands under the target. At an 8-hour day that is 620.34 ÷ 8 = $77.55 an hour, rounded up for the same reason.
The size of the mistake
The naive version of this sum was 60,000 ÷ 220 = $272.73. The arithmetic asks for $620.34, which is 620.34 ÷ 272.73 = 2.27 times as much. That gap is not a rounding error or a greedy markup. It is holiday, public holidays, sick days, the unbillable share, business costs, a pension and tax — every one of which a salary quietly absorbed and now does not.
The billable share moves the answer more than you expect
Hold everything else and change only that percentage. At 65%: 222 × 0.65 = 144.3 days, and 103,285.71 ÷ 144.3 = $715.78. At 85%: 222 × 0.85 = 188.7 days, and 103,285.71 ÷ 188.7 = $547.36.
Ten points down costs 715.78 − 620.34 = $95.44 a day. Ten points up saves 620.34 − 547.36 = $72.98. The curve is not symmetric — it steepens as the ratio falls — which is exactly why a bad quarter for admin hurts more than a good one helps.
Two different questions
What a target requires and what a client will pay are not the same question, and this arithmetic only answers the first. If the required rate is more than anyone will pay, that is genuinely useful to discover from a sum in an afternoon rather than from a year of invoices. It tells you which input has to move: the target, the costs, the days worked, the holiday, or the billable share.
Nothing here is a recommendation about price. No rate, ratio, holiday allowance or tax figure is built in, because all of them differ by country, by trade and by person — every number above came from the example inputs, and they are invented.
The Freelance Day Rate tool runs exactly this chain, shows the day-by-day derivation, gives a breakdown from rate down to take-home that foots to the cent, and runs it backwards: what a rate you can actually charge would leave, and how many billable days it would take to reach the target.