How to Ask for a Raise
Timing, the ROI case, the script, and what to do when the answer is no.
Most raise requests fail for a reason that has nothing to do with whether the person deserved it. They fail because the request was aimed at the wrong person. Your manager almost never controls the pay budget. They argue for it, in a meeting you are not in, against other managers arguing for their own people. Your real job is not to convince your manager you deserve more. It is to hand them something they can carry into that room and win with.
Why good people get turned down
The gap is almost never merit. It is usually evidence, timing, or who you were talking to.
Think about what happens after you leave your manager's office. In most companies of any size, they do not go and approve your raise. They add your name to a list. Weeks later that list meets other lists in a compensation planning meeting, where a fixed pool of money gets divided between more names than the pool covers. Someone has to be argued for. Someone has to be dropped.
Your manager walks into that meeting carrying whatever you gave them. If all you gave them was "I have been here three years and I work really hard," they have nothing to say when a peer manager says "so does mine." They lose the argument, come back, and tell you it was not the right time. That sentence is usually true. It is just not the whole truth.
The four failure modes
- Asked the wrong month. The money was already allocated eight weeks before the conversation happened.
- Brought effort, not outcomes. Hours, loyalty, and attitude do not survive a comparison meeting. Numbers do.
- Made it about need. Rent went up, a baby arrived, a mortgage renewed. All real. None of it is an argument a manager can repeat upstream.
- Left without a next step. "No" ended the conversation instead of starting the one that gets you paid in two quarters.
Every one of these is fixable, and three of the four are fixable before you ever open your mouth.
Timing: find the budget calendar
The single highest-leverage thing you can do costs nothing and takes one question.
Compensation runs on a calendar. Somewhere in your company there is a date when managers submit their proposed increases, and a date before that when the pool is set. Ask about it before you ask for anything else. Ask early enough that your name is on the list when the list is being written, not after.
The right window is roughly six to ten weeks before submissions are due. That is late enough that your recent work is fresh, and early enough that your manager can shape their proposal around you rather than squeeze you into what is left.
I want to make sure I bring this up at a useful point rather than at the wrong time. When does comp planning actually happen here, and when do you have to have your proposals in?
That one question is not a request, so it costs you nothing. It also tells your manager that you are thinking about the process the way they do, which is a small credibility deposit you will spend later.
Moments that raise your odds
- Right after you shipped something with a visible, attributable result.
- When your scope has just grown — a new report, a new account, a system you now own.
- When you have just been asked to take on more. This is the strongest moment there is, and it is fleeting.
- When someone left and you absorbed their work. Say so before it becomes the new normal.
Moments that lower them
- Immediately after a layoff, a missed quarter, or a hiring freeze announcement.
- The week your manager is buried in something else. You will get a distracted no.
- After the submission deadline has passed. You are now asking for an exception, which is a much bigger ask.
Build the case your manager can carry
Three kinds of evidence. Most people bring only the weakest one.
Evidence for a raise comes in three tiers. The higher the tier, the less it depends on anyone liking you.
| Tier | What it is | How it lands upstream |
|---|---|---|
| Money | Revenue won, cost removed, time saved that converts to money, churn prevented | Strongest. Survives a room full of strangers. |
| Scope | You now do meaningfully more, or harder, work than the job you were hired into | Strong. Reframes the ask as correcting a mismatch, not granting a favour. |
| Reliability | You are the one they hand the difficult thing to. Things stop breaking around you. | Real but weak alone. Needs one of the above beside it. |
Turn a win into an evidence line
A win in your head sounds like "I fixed the onboarding process." An evidence line sounds like this:
[The situation before] → [what I did] → [the measured change] → [what it is worth] "Onboarding took new hires about six weeks to get productive. I rebuilt the first-week sequence and wrote the runbook. It is down to about three weeks across the last five hires. That's roughly half a month of salaried output recovered per hire, and we hired nine people last year."
How many wins
Three to five. Not twelve. A long list reads as padding and forces your manager to pick the strongest one themselves — which means they might pick wrong. Choose the three that a person outside your team would understand without context, and lead with the one that has a dollar sign in it.
Set the number
Name one. A range invites them to pick the bottom of it.
Anchor on two things: what the market pays for the work you are actually doing now, and what the internal band for that work looks like. The second is often more decisive and almost nobody asks for it.
- Find the external number. Levels.fyi, Glassdoor, industry salary surveys, and — best of all — recruiters who have recently placed your exact role in your city. Recruiters will tell you the real range if you ask directly.
- Find the internal band. Ask your manager or HR: "What is the range for this level, and where in it do I sit?" Many companies will answer. In several jurisdictions, including a growing number of Canadian and US ones, they are required to.
- Pick a single number slightly above your target. Not a range. Ranges get read from the bottom up.
- Sanity-check it. A jump beyond roughly 15% usually needs a level change to justify it. If your number needs a promotion to make sense, ask for the promotion.
The conversation
Under a minute of talking, then stop and let them respond.
Book it as its own meeting. Do not bolt it onto a one-on-one where it competes with status updates. Twenty minutes, titled something plain like "Comp conversation" so nobody is ambushed.
I want to talk about my compensation, and I want to make it easy for you to take this upstream. Over the last year I've [win one, with the number]. I've also [win two]. And since [month] I've been [the scope that grew]. Based on what the market pays for this work and where I think I sit in our band, I'm asking to move to [number]. I know you don't decide this alone. Tell me what you'd need from me to make that case.
That last line is the important one. It moves you from petitioner to collaborator, and it gets your manager describing the constraints they are actually working under — which is the information you need.
After the pause
- Let the silence sit. The urge to fill it with justification is strong. Filling it is how people negotiate against themselves.
- Take notes visibly. It signals you intend to follow up, which changes how carefully people answer.
- Leave with a date. Never end on "let me think about it" with no calendar attached.
Put it in writing within 24 hours
This is the step most people skip, and it is the one that does the work. The email is what your manager forwards, pastes, and reads from when your name comes up. Make it short enough to be forwarded whole.
Subject: Following up — comp conversation Hi [name], Thanks for the time today. Putting the main points in one place so it's easy to reference. What I've delivered this year: • [win one — with the number] • [win two — with the number] • [scope that has grown since the role was set] What I'm asking for: [number], effective [cycle]. You mentioned [the constraint they named]. Happy to pull together anything that helps. [name]
The four things they push back with
These are near-universal. Have the answer ready and you stay in the conversation instead of retreating from it.
"There's no budget."
Usually true and usually not permanent. Do not argue with it. Convert it into a date and a condition.
That's fair, and I'd rather work with the real constraint than around it. When does the next budget open? And if I'm holding [the specific results] by then, is this the number we'd be working toward?
"Everyone gets the same increase."
This is a statement about the standard cycle, not about every mechanism that exists. Off-cycle adjustments, market corrections, and level changes all sit outside it.
Understood on the annual cycle. My question is more about the band than the increase. If the work I'm doing now maps to a higher level than my title, is a re-level the right conversation instead?
"You're already at the top of your band."
This is not a rejection. It is them telling you that more money requires a different job. Take that literally.
That's useful to know. Then what I'm really asking about is the next level. What would have to be true for me to be a credible candidate for it, and how long does that usually take from here?
"Let's revisit this later."
The most common answer and the most dangerous, because it feels like progress. Without a date and a written condition, "later" is a polite no with a long tail.
Happy to. Can we put a date on it now — say [specific month] — and can you tell me what you'd need to see between now and then? I'll send a short note capturing what we agree so we're both working from the same thing.
When the answer is no
A no is information. What you do in the next fifteen minutes decides what it costs you.
There are three kinds of no, and they call for different responses. Work out which one you got before you decide anything.
| The no | What it sounds like | What to do |
|---|---|---|
| Timing no | Budget is set, cycle is closed, freeze is on | Get a date and written conditions. Re-ask on schedule. This one usually converts. |
| Evidence no | They cannot see what you have done, or the work is invisible to their peers | Fix the visibility problem, not the work. Re-ask in one to two quarters with instrumented results. |
| Ceiling no | Top of band, no path described, vague answers repeated twice | Believe it. Your next raise is at another company or in another role. Start looking without announcing it. |
The 90-day agreement
Whatever kind of no you got, do not leave without this. It is short, specific, and written by you.
Subject: What we agreed — [month] check-in Hi [name], Capturing what we landed on so we're aligned. Between now and [date], I'll deliver: 1. [specific, measurable thing] 2. [specific, measurable thing] If those land, we'll revisit compensation at [date], targeting [number]. Let me know if I've mis-stated anything. [name]
Do not do these things
- Do not quietly disengage. Working visibly less after a no confirms the decision was right, and it is noticed faster than people think.
- Do not go around your manager on the first no. Escalating past them turns a solvable problem into a relationship problem.
- Do not re-ask in three weeks. Nothing has changed and it reads as pressure. One quarter minimum.
- Do not stop collecting evidence. The next conversation is already being built.