Mainstay

Guides

Guides

What a freelancer should cost

You came here for a number. There isn’t one, and the pages that give you one are guessing — usually without saying who they asked, when, or how. This page explains why the number cannot exist, what actually determines a price, and how to get a real figure for your own job in about a week.

Why no single figure describes anything

A rate is not a property of a skill. It is a property of a piece of work, its uncertainty, and who carries the risk of that uncertainty. The same person quotes differently for two jobs with the same job title, correctly, and a figure attached to the title averages over the thing that was actually deciding the price.

And the population is not one population. Independent professionals work in labour markets that have nothing to do with one another — different costs of living, different local demand, different currencies moving against each other. An average taken across them is a mixture: a number that describes no one who could be hired. Splitting it by country produces a cheapest-country table, which is a worse artefact than no table, because it invites you to shop for the country instead of the person.

Be sceptical of any figure that does not say who was asked, when, and how many. Most published rate ranges are assembled from other published rate ranges. When you find one that names its method and its date, it is worth reading — and it is still not about your job.

The eight things that actually move a price

Read these as the questions a good freelancer is silently pricing while you talk.

  1. How certain the scope is. Ambiguity gets priced. A brief that could mean two things is quoted against the more expensive one, or quoted low and repaired later at your cost.
  2. Who carries the revisions. “Two rounds included, then hourly” and “until you’re happy” are different products at different prices, and the second one is only offered by people who have been paid to learn what it costs.
  3. Whether you are buying hours or an outcome. An outcome transfers risk to them and they will price the risk. Hours transfer it to you, and hours are cheaper per unit for exactly that reason.
  4. Ramp-up. The first days of any engagement are paid learning about your product, your files and your people. On a short job that cost is a large share of the job, which is why small jobs have a floor that looks disproportionate and is not.
  5. How fast you decide. Clients who answer in an hour are cheaper to serve than clients who answer in a week. The difference reaches your quote whether or not anybody mentions it.
  6. The tail. What happens after delivery — fixes, questions, the handover to whoever maintains it. Priced if you ask for it, assumed if you don’t, and arguing about it afterwards is the most common way a good engagement ends badly.
  7. Your payment terms and your reputation for meeting them. Long terms are a loan and get priced like one.
  8. Whether they want the work. Interesting work with a competent client is discounted, quietly and universally. The clearest brief in the pile gets the best price, and nobody ever says so out loud.

A procedure that produces a real number

About a week, and it costs you one afternoon of writing.

Write one brief. How to brief and trial a freelancer covers the six things it has to contain. It has to be the same brief for everyone or the exercise is worthless.

Send it to three people and ask for a fixed price with stated assumptions. Not a rate. A price for this job, and the list of things they assumed to arrive at it. The assumptions are where the information is.

Compare the questions before you compare the prices. The person who asks what happens if the data is messy, who owns the final files, and who signs off, has priced a different and more realistic job than the person who replied with a number in ten minutes. A quote with no questions attached is a guess, and you will pay for the guess later, in scope arguments.

Read the spread as a measurement of your brief. If the three prices are far apart, that usually means the brief could be read several ways — the spread is measuring your ambiguity, not the market. Rewrite the brief with the assumptions you just learned about and ask again. The second spread is usually tight, and that number is your real price.

Then buy a small piece of it before committing to the whole, at the price agreed.

Hourly, fixed, or retainer — when each is honest

  • Hourly is honest when the scope genuinely cannot be known: discovery, debugging somebody else’s system, ongoing maintenance. You carry the risk, so you should also get a cap and a check-in, and you should expect an estimate even though it is not a promise.
  • Fixed is honest when the brief is tight and the outcome is describable. They carry the risk and will price it, so a fixed price is properly higher than the hours it looks like — that difference is not a markup, it is the risk transfer you asked for.
  • Retainer is honest when what you are buying is availability rather than output. It should state what happens to unused time, because that is where every retainer argument comes from.

The failure is using one where another was honest: fixed-pricing an unknowable scope produces somebody protecting their margin against you, and hourly-billing a well-defined deliverable makes speed a punishment.

The cheap-hire arithmetic

A low hourly rate is a price for hours. What you actually buy is the number of hours, and the number of hours is decided by how fast somebody understands your problem — which is not correlated with their rate in the direction people assume, in either direction. The way out is not to hunt for a lower rate; it is to ask for a fixed price on a defined outcome. Then the rate stops being the variable and the thing you are comparing is the thing you wanted.

What this site publishes, and what it will not

Every rate on this register is stated by the Mainstayer whose record it is. It is never verified, and it never can be here, because no money passes through this site — there is no escrow, no invoicing and no commission; you pay the person directly.

And no benchmark table will be published here. Not a median, not a range, not a “typical rate” by skill. Any figure of that kind computed from this register would describe this register rather than the market.

What is published instead is the bar every record was read against — the standard, numbered and dated — and, on what a directory cannot verify, the full list of things nobody here can check for you.

Open the register →

Browsing is free · a card is needed only to reveal a contact