Pricing, from first principles to the annual review

Price rating work by making seven decisions in order: measure your time, choose the unit, build tiers, price each format, work back from take-home, position inside the market band, and hold the number in conversation. Most pricing problems come from making a later decision before an earlier one.

The usual versions are setting a number before measuring the time it costs, or negotiating before having a card to negotiate against.

This is the whole sequence, at summary depth, with the post that treats each part properly linked from its section. It is not a substitute for those; it is the order to read them in and the argument that connects them.

One thing it deliberately does not cover: what the trade pays on average. That is a different question with a different answer, and what rating work actually pays handles it. This is about what you charge, which you control.

Decision one: measure the time before you price anything

You cannot price an hour you have never measured, and every earner's estimate of their own minutes per job is wrong in the same direction.

Do one complete job on practice material and time all of it: reading the brief, forming the judgement, producing the deliverable, sending it, and the admin afterwards. Then add the unbilled time that surrounds it - the enquiries that went nowhere, the clarifying questions, the being-reachable. The practice run is the method, and the honest version routinely lands at roughly double what people assumed.

Two properties of that number matter more than its value.

It falls, sharply, over the first couple of months, which means your effective hourly rate rises without you touching your price - why your effective hourly rate keeps changing works through the mechanism. And it is not the same across formats, which is why a single price across a menu is nearly always wrong on at least one line.

The unbilled surround is the part that decides whether this is worth doing at all. The arithmetic of a part-time week puts a figure on it: roughly a third of a session goes before any deliverable is produced, and pricing that ignores it prices a fiction.

Decision two: choose the unit

Per job or per hour. This is not administrative, and it decides who carries the risk of the work running long.

Per-job pricing puts that risk on you and is what this trade prices in almost everywhere, because buyers want a number before they commit and a clock makes them nervous. Hourly pushes the risk onto the buyer and is defensible for open-ended or live work. Hourly versus per-job pricing is the full argument.

The consequence of choosing per job, which is the choice nearly everyone makes, is that scope becomes a pricing instrument rather than a formality. Every extra twenty minutes you agree to is a discount you did not decide to give, which is why time-boxing a job sits inside the pricing question rather than beside it.

Decision three: build tiers, not a price

A single price answers one buyer and loses the two either side of them.

Three tiers is the working structure, and the top one earns its place even if it never sells - it makes the middle read as the sensible choice, which is the entire mechanism described in why your top tier matters unsold. Its price works as the anchor in the sense Tversky and Kahneman (1974, Science) described, where estimates adjust from whatever relevant value is available. The bottom tier does the opposite job and carries a cost worth knowing about: the cheapest tier reliably attracts the buyers who ask for most, which is adverse selection, the mechanism Akerlof (1970, Quarterly Journal of Economics) described in markets where one side cannot judge quality in advance.

The output of this decision is a rate card, and the point of the card is not to publish prices. It is to make every individual pricing conversation already settled before it starts, which is the argument in setting a rate card you can defend.

Decision four: price the format, not the effort

Format premiums in this trade are wider than the effort gap explains, and the reason is supply rather than quality.

Recorded work prices above written work by a margin that has little to do with the extra minutes, because far fewer people are willing to record - what a recorded review adds to the price covers that gap and where it comes from. Audio sits close to video rates with none of the lighting, framing or editing, which makes it the most consistently underpriced line on most cards; audio only, as an underpriced format is the case for putting it on yours. Live work prices above everything recorded for a structural reason: a booked hour cannot be resold, and that is what the price of a fixed slot is really charging for.

The same observations organised into defined axes also sell for more than the same observations written as prose, which is a formatting decision with a price attached - written versus scored has the comparison.

Urgency is its own line. A same-day turnaround costs you the ability to schedule your own day, and that, rather than the speed, is the thing a rush fee is charging for; rush fees and what urgency is worth sets out how to structure one so it does not become the default.

Decision five: work backwards from take-home

The price is not the money. Four separate deductions sit between a buyer's payment and your bank account, and only one of them is usually advertised - how platform fees change your effective rate enumerates them.

Two of those deductions do not scale down, which has a direct consequence for the bottom of your card. Flat processing charges make the smallest job the one most likely to be underwater, and the minimum viable price when fees are flat is the arithmetic that finds your floor.

If your buyers, the platform and your bank are not all in one currency, there are two conversions in the chain that you did not choose and are paying for anyway - currency and cross-border payouts covers what that costs and where it hides.

Set the floor from take-home, then price up from it. Pricing down from a headline number is how a card ends up with a tier that loses money on every sale, discovered eleven months later during a reconciliation.

Decision six: locate the band, then position inside it

You are not pricing in a vacuum, and you are also not pricing against the three profiles you happened to look at.

Three profiles is not a market, and the three you looked at were the ones ranked highest, which is a biased sample in a specific direction - reading the market band without guessing is the sampling method that fixes it.

Position inside the band rather than at the bottom of it. Nearly all of the gap between two earners charging very different rates for what looks like identical work is structural rather than qualitative, and why two earners charge very different rates itemises what those structures are. Undercutting the band communicates inexperience, which is the one thing a new earner is trying not to communicate, and it selects for the buyers most likely to leave the moment you correct it.

The floor of this market moved for a reason worth understanding rather than resenting. Generic assessment is now free and instant, so anything on your card that a model could produce is already priced as though a model produced it - the technical account of what automated scoring reliably does is the clearest map of which lines are exposed, and the consumer tools built on it are what a buyer has usually already tried before they come looking for a person. What survives at a price is judgement by a named individual who actually looked, and that is the thing to be selling.

Crowding, when it arrives, does not lower every price uniformly. It lowers the price of whatever is easiest to supply, which is the mechanism in what happens to rates when a market gets crowded and the reason a scarce format is worth more than a good one.

Decision seven: hold the number in conversation

A card only works if it survives contact with a buyer.

"Can you do it cheaper" has a correct answer that is not yes and not no, and answering it is a technique rather than a temperament. Every job has four variables and the buyer only ever names one, which is what makes negotiating without lowering the number possible: scope, turnaround and format all move before price does.

When you do discount, discount structurally rather than on identical work. A smaller job at a lower tier keeps the card intact; the same job at a lower price retires it - the distinction is the whole of discounting without wrecking your rate card. Bundles are the honest form of this, because repeat work genuinely costs almost nothing to acquire and that saving is real rather than invented; bundles and repeat buyer pricing has the reasoning.

Publishing the number at all is itself a decision. Quote-on-request removes everybody who was not going to negotiate, which is nearly everybody, and publishing prices versus quoting on request works through when the trade is worth making.

The rise, and the review that schedules it

A price rise with no queue behind it makes the queue longer in the wrong direction. The evidence that justifies one is a backlog, not a feeling - the first rate rise sets out the trigger and the size, and waiting lists and what they signal explains why a backlog is the only real proof that your rate is under what the market will bear.

Existing buyers need handling separately from the listing. A repeat buyer discovering a rise from a price page is a worse conversation than one you opened yourself, which is the case for grandfathering old buyers deliberately rather than by accident.

And when bookings go quiet afterwards, resist the reversal. A fortnight of silence after a change is usually a fortnight; when bookings stop after a rise covers how long to wait and what would actually constitute evidence.

Then put it on a schedule. Rate cards drift into whatever the first version was, and the first version was a guess made before you had measured anything - the annual review of your own rate card is the hour that stops a year-old guess becoming a permanent one.

The one number that outranks all of it

None of this matters if the result loses to your alternative. A rate that beats the market band and pays less per hour than the work available to you next month is still the wrong rate, and what your time is worth outside this trade is the comparison that settles it.

Two things sit outside this site entirely and are worth knowing where to find. The measurement conventions and their precision, if precision is what you are selling, are documented properly elsewhere rather than here. And the platform-specific side - what a given marketplace's tiers, fees and listing structure actually allow you to charge - is the platform's own to publish, which for this one is Rate Cock's judges page.