Pay
What rating work actually pays
The honest answer is a range with conditions attached, and the conditions matter more than the range does.
Anyone who gives you a single number for what rating work pays is either selling something or has done it exactly once. The work is priced per job, the jobs vary enormously in length, and the same person can earn four times as much in November as they did in September without changing anything about how they work.
What follows is the structure of the number rather than the number itself: the inputs that decide what an hour of this is worth, in roughly the order they matter. Once you can see the structure, a rate card stops being a guess.
The unit is a job, not an hour
Almost nobody in this trade is paid by the hour, and that single fact drives most of the confusion about what it pays.
A buyer commissions a thing: a written assessment, a scored breakdown, a short recorded reaction, a comparison across several submissions. They pay a price for that thing. How long it takes you is your problem, which is a bad deal on your first ten jobs and an increasingly good one after that.
This is why "what does it pay per hour" is answerable only backwards, after the fact, by dividing. It is also why the biggest single lever on your effective rate is not price at all. It is how long the work takes you, which is a skill that improves faster than your rate does. A job that took forty minutes in your first month and takes twelve in your fourth has tripled your hourly income at an unchanged price.
The five things that actually move the number
In descending order of how much they matter:
| Input | Effect on what you earn | How fast you can change it |
|---|---|---|
| Time per job | Direct multiplier on hourly rate | Weeks - this is practice |
| Volume of jobs | Direct multiplier on monthly total | Months - this is reputation |
| Price per job | Direct, but capped by the market | Slowly, and only with evidence |
| Format mix | Recorded work prices well above written | Immediately, if you can do it |
| Platform cut and payout costs | Silent 10-40% reduction | Once, by choosing carefully |
Notice that price is third. New earners spend most of their attention there because it is the number they control most obviously, and it is the one with the least room in it. The market sets a fairly narrow band for a given format, and being 20% above the band mostly buys you silence.
Format is the biggest jump available on day one
Written work is the floor of this market almost everywhere. It is the easiest to produce, the easiest to compare, and therefore the most price-competitive.
Recorded work - audio or video - prices substantially above written work for the same subject, and the gap is wider than the extra effort justifies once you are practised. The reason is not mysterious: fewer people are willing to do it, and it cannot be quietly copied from a template. If you are comfortable on camera or on a microphone, that comfort is worth more than any pricing tactic in this article.
Structured or scored work sits in between. A rating with an actual rubric behind it - defined axes, a consistent scale, a written justification per axis - reads as a professional product rather than an opinion, and it prices accordingly. The mechanics of how a scored rating is put together are a subject in themselves, and Penis Rater covers how scoring systems are built better than a post about money can.
What a realistic month looks like
Any figure here is an illustration, not a promise, and your country, platform, format and hours make it move. But the shape is consistent enough to be useful.
A first month is quiet. Nearly everyone is surprised by this, and nearly everyone quits during it. You have no reviews, no completed jobs, and no reason for a buyer to pick you over someone with thirty of each. Most of the first month's value is not income, it is the small number of completed jobs that make month two possible. The mechanics of getting out of that hole are covered in the post on the cold-start problem.
Months two and three are where the trade either becomes real or does not. Volume arrives before price does: you get more work at the rate you started at, and your hourly income rises because you got faster, not because you got dearer.
By month four or five, if the work is still coming, you have the two things a rate rise requires - a completion record and a queue. Neither argument works without the other. Raising your price with no queue just makes the queue longer in the wrong direction.
The cut nobody budgets for
Your headline price is not your income, and the gap between them is bigger than most new earners assume.
Between what a buyer pays and what lands in your account there is typically some combination of a platform commission, a payment processor fee, a currency conversion spread if you and the platform are not in the same currency, a payout fee on withdrawal, and a minimum withdrawal threshold that quietly holds a float of your money hostage.
None of these are scandalous individually. Together they are the difference between a rate that works and one that does not, and they are why two earners charging the same headline price can take home noticeably different amounts. The post on platform fees works through how to calculate the version of your rate that actually matters.
Check the withdrawal threshold before you check the commission. A 15% commission with a $20 threshold is better for a beginner than a 10% commission with a $200 one, because in month one the second option pays you nothing at all.
Where the demand comes from, and why it is lumpy
Demand for rating work is not evenly spread across the week, the month, or the year. It clusters in evenings and weekends, it rises around the paydays of whatever countries your buyers are in, and it moves with whatever is currently driving people to think about the subject at all.
This lumpiness has a practical consequence: your availability during the busy hours is worth more than your availability in total. Twelve hours spread across the quiet parts of a week can earn less than four hours placed correctly.
It also means a bad week is usually a bad week rather than a signal. The trade punishes people who reprice in response to a fortnight of noise.
The parts of the job that are not the job
The thing you are paid for is the assessment. The things you are not paid for, and which nonetheless consume your hours, are:
- reading and clarifying the brief
- declining work that does not fit
- chasing a buyer who has gone quiet mid-job
- handling a revision request that was not in scope
- the admin of getting paid, and the records that go with it
Count these when you work out what an hour is worth. A job priced well against its production time can be priced badly against its total time, and the difference is entirely made of small unbilled things.
There is also the work you do that never becomes a job at all. Enquiries that go nowhere, briefs that turn out to be someone shopping around, messages that arrive at two in the morning and are gone by breakfast. Some fraction of your attention is spent on demand that never converts, and that fraction is a real cost even though nothing about it appears on an invoice. Earners who track it are usually surprised by how large it is in month one and how quickly it shrinks once a profile has a record behind it.
Scope creep is the largest of them. A written assessment that turns into a conversation, then a follow-up, then a second look at a new photo, is three jobs sold at the price of one. The fix is a scope sentence in your own listing, not a confrontation later.
AI changed the floor, not the ceiling
Automated scoring is now cheap and instant, which has taken the bottom out of the market for anything that reads as a generic verdict. If what you produce could be produced by a model in two seconds, it will be priced as though it were.
What it has not touched is the top of the market, and in some ways it has made the top more valuable by making the contrast obvious. People commission a person for judgement, specificity, and the fact that a human being actually looked. The technical account of what automated scoring is doing is worth reading precisely because it makes clear what a model cannot supply.
The practical implication for pricing is to stop competing on the axis where you lose. Speed and volume belong to the machines now. Specificity, judgement and being a named person with a track record do not.
What to do with all this
Work out your time per job before you work out your price. Then calculate your effective rate after every deduction, not the headline one. Then leave the price alone for two months while you get faster, because that is the lever with the most slack in it.
If you want to see the demand side of the same market - what buyers are actually asking for and how they choose - Rate Penis covers commissioning from the buyer's position, and the gap between what they want and what earners assume they want is where most pricing mistakes live. For the platform-specific version of any of this, including how to start taking paid work on the service this site is published by, Rate Cock's judges page is the place with the current numbers.
The trade rewards patience in an unusually literal way: the same job, done the same way, is worth more to you in month five than it was in month one, and nothing about that requires anyone to pay you more.