One sheet answers the only pricing question that matters

Take-home per hour, computed separately for each format you offer, is the single number that decides your rate card, and almost nobody calculates it. The reason is that it requires two things people record inconsistently: what actually landed in the account, and how long the job actually took.

Thirty jobs is roughly the point where the answer stops being noise, and the answer is usually that one of your formats is paying half of what another is.

The five columns

You need one row per job and five columns. Anything more is a project you will abandon in week three.

Format. The label you sell it under, not a description. Price. What the buyer paid. Take-home. What reached your account after every deduction, including conversion spread. Those deductions are not small: Stripe's published UK pricing is 1.5% + 20p for standard UK cards, 3.15% + 20p for non-EEA cards, plus 2% when currency conversion is needed. Minutes. Reading the brief, doing the work, delivering, and the admin after. All of it. Outcome. Completed, revised, refunded.

That last column is the one people skip and it is the one that separates a format that pays well from a format that pays well when it goes right. The rest of the record-keeping case is in the post on records; this is the subset that answers a pricing question rather than a tax one.

The arithmetic

Take-home divided by minutes, multiplied by sixty. Averaged per format.

Then two adjustments that change the ranking more often than not.

Subtract refunded jobs' time without adding their fee, because that time was spent. And add the unbilled minutes each format costs outside the job itself: enquiries that did not convert, revision negotiation, scheduling.

An illustration, not market rates. The shape is the point: the highest-priced format is the worst-paying one, which is the usual result and the reason price alone tells you nothing. Comparison work prices badly for structural reasons set out in the post on comparison jobs, and this is what that looks like in a spreadsheet.

Read the spread before the average

Two formats can average the same and behave completely differently.

A format with a tight spread is predictable and can be priced close to its average. A format with a wide spread contains jobs that take three times the median, and its price has to cover the bad tail rather than the middle, which usually means it needs to be more expensive than the average suggests.

Look at your slowest three jobs in each format. If they are all the same format, that format has a scope problem rather than a pricing problem, and the fix is the listing text rather than the number.

Acting on the result

Four moves, in descending order of how often they are the right one.

Raise the price of the slow format. If the work is worth doing and the rate is wrong, correct the rate. The first rate rise covers the timing.

Tighten its scope. Frequently the format is fine and the version you have been delivering is larger than the one you listed.

Reprice by tier rather than dropping it. A slow format at a high price selects for buyers who want it specifically, which is a smaller and better market.

Stop offering it. The right answer when it is slow, tightly scoped already, and priced at the ceiling the market will bear. When to stop taking a format is the longer version, including the reasons people keep loss-making formats on the card for years.

Do one of these at a time and leave it a month. Changing price and scope together means you learn nothing from what happens next, which is the same discipline as price testing without annoying anyone.

The number moves, and that is the useful part

Your per-hour figure on any format improves markedly in the first few months as the work stops being novel. This means an early measurement will understate a format you are still learning, and it is a real argument against dropping something on the first thirty jobs.

Recompute quarterly. A format that is still bottom of the table after two quarters is not a learning curve, it is a structural mismatch between what buyers will pay and what the work costs you.

Watch the ordering rather than the absolute figures. Absolute rates move with your prices, your speed and the market; the ranking between your formats is stable and is what the decision actually rests on. Why the effective hourly rate keeps changing covers the drift, and this is the version of it you can act on.

What the sheet cannot tell you

It cannot tell you which formats bring buyers back, and repeat work has an acquisition cost of near zero that never shows up in a per-job calculation. A format that pays badly and produces repeat buyers may be the most profitable thing you offer once the second and third bookings are counted against it.

If your records include buyer identifiers, add a sixth column for whether the job was a first or a repeat, and the picture changes again.

It also cannot tell you what buyers want next, which is a demand question rather than a records one. The buyer-side account of how a format gets chosen is set out from the commissioning side, and it explains several of the rankings a spreadsheet only describes. The formats under most pressure are the ones closest to what an automated tool already produces free, which the technical account of automated scoring maps precisely. Where your format's value rests on being methodologically careful, the conventions worth stating explicitly are what make that legible to a buyer.

For what actually gets deducted before the take-home column - the input the whole calculation depends on - Rate Cock's payouts page is where the current figures live.