How to price your freelance work (and stop undercharging)

For my first two years freelancing I charged by the hour, and I got steadily better at the work. My income went down. Not dramatically, but measurably, because a job that used to take me six hours took four, and I dutifully billed four.

I was being punished for competence, by a system I’d chosen myself. And that was before AI made everything faster again. If you’re billing hourly in 2026, you’ve built a business that shrinks every time you improve.

Why hourly billing is now actively harmful

The maths has always been awkward and AI has made it absurd.

A task that took you a day now takes three hours. Under hourly billing, you just took a 60% pay cut for learning a better method. The client’s outcome is identical, the value they received is identical, and your income fell.

Meanwhile the freelancer who hasn’t improved bills the full day and earns more. The system rewards exactly the wrong thing.

There’s a second, quieter cost: hourly billing invites clients to scrutinise your hours rather than your results. Conversations become about timesheets instead of outcomes, which is a worse conversation for everyone.

What to charge instead

Price the project, based on what the outcome is worth to the client.

Those two numbers, your time and their value, have never been related, and pretending otherwise is what keeps freelancers underpaid. A landing page that takes you six hours might generate significant revenue for a business. Your six hours is not the relevant figure.

The questions that reveal the value, asked on the call:

• What happens if this doesn’t get done? The cost of inaction is the ceiling on what they’ll pay.
• What’s this worth to you if it works? Sometimes they’ll tell you a number. It’s usually larger than yours.
• Who else is involved and what’s the deadline? Urgency and visibility both raise value.
• Have you paid for something like this before? Anchors you to their reference point.

Record the call so you can review the answers properly. AI meeting notes tools that actually save you the follow-up covers the tooling, and this is one of the higher-value uses of it.

Building the number

Practical method, once you know roughly what the outcome is worth:

Start with your floor. What you need per month, divided by realistic billable capacity. Not 40 hours a week, more like 20 once you account for admin, sales and the rest. This is a floor, not a price.

Estimate the project honestly, then add 30%. Every freelancer underestimates. Every one. The 30% is not padding, it’s the correction for a bias you definitely have.

Sanity-check against value. If your number is a tiny fraction of what the outcome is worth, raise it. If it exceeds what the outcome is worth, this isn’t a project worth taking.

Offer three options where it makes sense. A reduced scope, the recommended one, and an expanded one. This shifts the conversation from whether to which, and a surprising number of people pick the expensive one.

Put it in a proposal the same day. AI proposal and contract tools for freelancers who hate paperwork covers making that fast, and speed genuinely closes deals.

Where AI helps with pricing

Less than you’d hope, and in specific places.

It’s good at: structuring the options, drafting the pricing section, playing the client so you can rehearse the awkward conversation, and analysing your past projects to find where you consistently underestimate. That last one is worth doing: feed it your last twenty projects with quoted price and actual hours, and the pattern is usually humbling.

It’s bad at: telling you what to charge. It has no idea what your market pays, what your reputation is worth, or how badly this particular client needs the work done. Any number it gives you is a plausible-sounding average, and averages are what you’re trying to escape.

The signal that you’re too cheap

Nobody flinching.

If every prospect says yes immediately, your price is below the point where they have to think about it. That feels great and it’s costing you money.

Healthy pricing produces some friction. Some negotiation, some lost deals. Roughly one in five saying no is about right. If you’re at zero, raise your rates on the next quote and watch what happens, which in my experience is usually nothing, because the number was never the obstacle you imagined.

Other tells: you’re fully booked and not earning enough, or you find yourself resenting reasonable clients. That resentment is almost always a pricing problem rather than a client problem.

Raising rates on existing clients

The conversation everyone avoids. What’s worked for me:

Give notice, roughly a month. Be brief and unapologetic: rates are going to X from this date. Don’t over-explain, don’t list justifications, and don’t apologise, because all three invite negotiation.

Expect to lose some. That’s the mechanism working, not failing. The clients who leave over a modest increase are usually the ones consuming disproportionate time.

Track who’s on what rate, which sounds obvious and is where things get messy after a few years. I keep this in a CRM alongside everything else, and HubSpot’s free tier handles it fine. My HubSpot free CRM review covers how far that stretches, and CRM-lite for solopreneurs has lighter options.

My take

Move off hourly billing. That single change matters more than any pricing technique, and it matters more now than it did three years ago, because the gap between how long work takes and how long it used to take keeps widening.

Then get comfortable with resistance. The freelancers I know who earn well aren’t better at the work than the ones who don’t, they’re just charging numbers that make some people say no. Everyone underprices at first, myself very much included, and the correction is almost always upward.

And ask the value questions on the call. Not to squeeze anyone, but because you genuinely cannot price something sensibly when you don’t know what it’s for. Half the time the client tells you a number bigger than the one you were about to say, and that conversation is worth more than any rate calculator. AI side hustles that actually make money covers the earlier stage, if you’re not at the pricing conversation yet.