TL;DR: Raising your rate with a client you already work for is a mechanical change plus a conversation, and most freelancers only get the mechanics right. On an hourly contract you propose a new rate from the contract page and the client has to accept it, after which it applies to hours logged from then on, never to work already billed. On fixed-price there is no standing rate at all, so the increase happens in the price you quote for the next milestone or the next project. The conversation is short: give notice, name the new number and the date it starts, point at what has changed since the contract began, and do not apologise or ask permission. Ask for a step you would actually accept, typically in the range of ten to twenty percent, and tie it to evidence rather than to inflation or need. The part nobody mentions is leverage: you can only raise a rate you are genuinely willing to lose, which means the real prerequisite is a pipeline of other work, not a better script. This guide covers when you have earned it, what the platform lets you change, the exact wording, and what to do when the answer is no.
Most freelancers on Upwork are underpaid by their own past selves. The rate on a contract is a number you chose months or years ago, when you knew less about the client, less about the work, and less about what the work was worth. The client is not going to revisit it, because nobody has ever woken up wanting to pay more. So the number sits there, quietly turning your best relationship into your worst-paid one.
The good news is that this is one of the few parts of freelancing where the awkwardness is much bigger than the risk. A rate conversation with a client who values you usually ends in a yes, or in a no that costs you nothing except the information that this client has a ceiling.
Why the raise never happens on its own
Two things keep rates frozen. The first is that Upwork contracts are open-ended by default. There is no renewal date, no annual review, no moment the system hands you to force the question. The contract just continues, and so does the rate.
The second is that the freelancer is the only person in the relationship who experiences the problem. From the client's side, nothing is wrong. You are delivering, the invoices are predictable, and the budget line has not moved. They are not withholding a raise, they are simply not thinking about one. Waiting for a natural moment means waiting for something that never arrives, which is why setting a rate that only applies to new proposals leaves your oldest clients paying your oldest prices.
When you have actually earned it
Ask when at least one of these is true, and say which one in the message.
The work has changed shape. You were hired to do one thing and are now doing a broader or more senior version of it: managing a process instead of executing a task, being consulted before decisions, owning an outcome rather than a deliverable. This is the strongest case, and it is also the honest version of the thing that would otherwise be scope creep.
Your evidence has improved. A stronger portfolio, a badge like Top Rated, completed contracts in a higher-value niche, a certification the work actually uses. The client is not paying for your time, they are paying for the confidence that the job gets done, and that confidence is now better founded.
Your market rate has moved. You are winning new contracts at a higher number than this client pays. That gap is the entire argument, and it is a fact rather than an opinion.
You are at capacity. When every hour you give one client is an hour taken from another, price is the only tool you have left. Uncomfortable to say out loud, completely legitimate to act on.
Enough time has passed. Twelve months of good delivery on the same contract is reason enough for a modest step. Nobody needs a dramatic justification for a rate that has not moved in a year.
What is not an argument: your costs went up, the platform takes a cut of your earnings, or your rent rose. All true, none of them the client's problem, and each one invites a negotiation about your circumstances rather than your value.
What the platform actually lets you change
The mechanics differ by contract type, and getting them wrong makes an otherwise good conversation look amateurish.
Hourly contracts. Open the contract from your Jobs list and propose a new hourly rate. The client receives the request and has to accept it before anything changes. Once accepted, the new rate applies to hours you log after that point, not to time already billed or to the week in progress. If the client ignores the request, the contract simply continues at the old rate, so treat a silent request as a no rather than as a pending yes.
One detail that catches people out: most hourly contracts carry a weekly hour limit, and the client sees the raise as an increase in their weekly maximum spend, not as a rate on a page. A twenty percent rate rise on a twenty hour limit is a twenty percent bigger invoice every week. If their budget is fixed, expect them to accept the rate and trim the hours, which is a perfectly reasonable outcome and one you should decide how you feel about in advance.
Fixed-price contracts. There is no standing rate to change. The price lives in each milestone, so a raise means quoting the next milestone or the next project higher, and new milestones need to be funded in escrow before you start. That makes the natural moment obvious: the end of the current scope, when you are scoping the next piece, which is one of the practical differences behind fixed-price versus hourly. Do not try to reprice work that is already funded and underway.
How much to ask for
Work from three numbers: your floor, your current public rate, and what this client pays today. A step you will actually go through with beats a perfect number you keep postponing. Ten to twenty percent is the range most freelancers can ask for without a long justification, and it can be repeated next year. Larger jumps are defensible, but they need a change in the work to explain them, such as taking on a responsibility the client would otherwise hire for.
Two rules keep the number honest. Never land below your floor, the take-home you calculated after unbillable time and the service fee, because a contract under your floor is a contract you are subsidising. And do not exceed the rate you quote to new clients, because it is a strange position to defend and the client can see your public rate on your profile.
The conversation, in order
Pick the moment. Right after a delivery that went well, never after a mistake, a missed deadline, or a difficult week. If something has just gone wrong, fix it, wait a month, then ask.
Give notice, do not spring it. Two to four weeks between the conversation and the new rate taking effect. That gives a client with an approval process time to use it, and it is the single thing that most separates a professional rate change from an ultimatum.
Say it in the message thread first, then send the platform request. A rate change request that arrives with no warning reads as a system notification rather than a conversation, and the easiest thing to do with a system notification is ignore it.
Keep it to one short message. Length reads as anxiety. Something close to this:
Hi [name], I wanted to give you plenty of notice that I am moving my rate to $X from [date]. Over the last [period] the work has grown from [original scope] into [what it actually is now], and that is the rate the same work goes out at for new clients. Everything else stays as it is, and I am happy to keep the weekly hours where they are or adjust them if the budget needs to stay flat. I will send the change through Upwork so you can approve it there. Glad to talk it through any time.
Then stop typing. Do not add a paragraph explaining that you understand if it is difficult, do not offer a discount before anyone has objected, and do not ask whether the increase is okay. You are informing a business partner of a price change with notice, which is an ordinary thing that ordinary businesses do.
When the answer is no
A no is information, not a verdict on you, and you have four moves.
Hold the scope to the budget. If they cannot pay more per hour, reduce what you cover so the lower rate matches a smaller job. This is often the fairest outcome and it stops a frozen rate from quietly becoming a bigger job.
Phase it. Half now, the rest in six months, agreed in writing. Useful when a client wants to pay you and has a budget cycle in the way.
Set a date. Agree a specific month to revisit, put it in your calendar, and raise it again then. A vague "let us look at it later" is a no with better manners.
Taper and replace. Keep the contract, stop giving it your best hours, and put the freed capacity into finding a replacement at the rate you want. When the replacement lands, end the contract cleanly with notice and a proper handover, because the review sits on your profile long after the money is spent and a bruised client is a real risk to your Job Success Score.
What almost never works is an immediate ultimatum. Clients who feel cornered comply and then quietly start looking, and you lose the contract anyway, on worse terms and with worse feedback.
The part that actually decides the outcome
You can only raise a rate you are willing to lose. That is the whole thing. Every script in this article is just wording on top of one question: if this client says no and leaves, what happens to your month?
Freelancers with a live pipeline ask calmly, accept a no without flinching, and get a yes surprisingly often, because clients can hear the difference. Freelancers with one client and no incoming work ask apologetically, discount before they are pushed, and usually do not ask at all. The fix is not a better message, it is having somewhere else to go, which makes a steady flow of genuinely good jobs the real prerequisite for every rate conversation you will ever have.
The obstacle is that maintaining a pipeline while fully booked is exactly when nobody has time to do it. The good posts collect most of their proposals in the first hours, and Upwork's own search filters will not narrow by client spend, hire rate, or rating, so staying in the market means opening posts one at a time between deliverables.
That is the job Upwork Scout was built to take over. It watches Upwork continuously, reads the full job detail rather than the listing, enforces your floors on client spend, hires, rating, payment verification, and budget before anything reaches you, then scores what survives against your actual profile. Fewer jobs arrive, they come from clients with a history of paying real money, and they arrive early enough that a careful proposal still gets read. The free tier lets you set a minimum budget floor today, which is the cheapest way to make sure your next rate conversation is one you can afford to lose.
Mistakes worth avoiding
Raising the rate on work in progress. Finish the funded milestone or the agreed sprint, then move the number. Repricing mid-delivery is the thing clients remember.
Asking by hint. "I was wondering if maybe there was any room to look at my rate at some point" is not a request, it is an invitation to say nothing.
Justifying with your costs. Value, evidence, scope. Never rent.
Raising it on your worst client first. Practise on a relationship that works. The difficult client is the one to replace, not the one to negotiate with.
The bottom line
Rates on existing Upwork contracts do not move unless you move them. Pick a reason that belongs to the work rather than to your circumstances, choose a step you will actually send, give two to four weeks of notice in the message thread, then submit the change through the contract so the platform side matches the conversation. Take a no gracefully and act on it. And keep enough new work coming in that the whole exchange stays what it should be: a routine price change between two businesses, rather than a request for a favour.
Frequently asked questions
Can you change your hourly rate on an active Upwork contract? Yes. Open the contract from your Jobs list and propose a new hourly rate, which the client then has to accept before it takes effect. The new rate applies only to hours you log after they approve it, so already-billed time and the current week are unaffected, and there is no way to apply an increase retroactively. If the client does not respond, the contract carries on at the old rate, so treat silence as a no and follow up in the message thread. On fixed-price contracts there is no rate field to change at all, because the price sits in each milestone, so the increase has to go into the next milestone or the next project you quote.
How much should you raise your rate with an existing client? For most freelancers a step of roughly ten to twenty percent is the sweet spot: large enough to be worth the conversation, small enough that it does not need an elaborate justification, and repeatable in a year. Bigger jumps are defensible when the work itself has changed, for example when you have taken on responsibility the client would otherwise have hired someone else for. Set two boundaries before you pick a number: never go below the floor you calculated from your own take-home requirement after the service fee, and do not go above the rate you quote to new clients, since that number is visible on your profile.
What if the client says no to a rate increase? Treat it as information rather than a rejection. You can hold the scope down so the smaller rate matches a smaller job, phase the increase over two steps, or agree a specific date to revisit it rather than a vague "later". If none of that works and the contract is well below your market rate, keep delivering professionally while you replace it, then end the contract with proper notice and a clean handover, because the review outlasts the contract. The one option to avoid is an immediate ultimatum, since clients who feel cornered tend to agree and then start looking for a replacement anyway.