Asking for a Raise: Timing, Script and Evidence

Asking for a raise feels like a single tense conversation. In most companies, it is really a process with a calendar: managers propose increases before a review cycle, budgets are fixed, and the meeting where you hear the result comes at the end. Knowing that changes when and how you ask.

What typical raises look like

Wage data gives a useful reference point. The Atlanta Fed's Wage Growth Tracker follows US workers a year apart and reports the median change in their hourly pay, separately for people who stayed in the same job and those who switched (Atlanta Fed: Wage Growth Tracker).

import pandas as pd

# Atlanta Fed Wage Growth Tracker (US Current Population Survey): median 12-month change in hourly pay
overall = pd.read_excel("wage-growth-data.xlsx", sheet_name="data_overall", header=1)
overall = overall.rename(columns={overall.columns[0]: "month"}).dropna(subset=["month"])
for col in ["Job Stayer", "Job Switcher"]:
    overall[col] = pd.to_numeric(overall[col], errors="coerce")
overall["year"] = overall.month.dt.year

print("W1 median wage growth of people who stayed in the same job (3-month moving average), yearly mean")
years = overall[overall.year >= 2019].groupby("year")[["Job Stayer", "Job Switcher"]].mean()
for year, r in years.iterrows():
    print(f"   {year}  stayers {r['Job Stayer']:.1f}% | switchers {r['Job Switcher']:.1f}%")

occ = pd.read_excel("wage-growth-data.xlsx", sheet_name="Occupation", header=2)
occ = occ.rename(columns={occ.columns[0]: "month"}).dropna(subset=["month"])
occ["Professional and management"] = pd.to_numeric(occ["Professional and management"], errors="coerce")
last = occ.dropna(subset=["Professional and management"]).iloc[-1]
print(f"\nW2 professional and management occupations, 12-month moving average to {last.month:%b %Y}: "
      f"{last['Professional and management']:.1f}% (all workers {pd.to_numeric(last['Overall']):.1f}%)")

print("\nW3 what a stayer's typical raise would mean: the same pay growth sustained for five years")
for g in (years.loc[years.index.max(), "Job Stayer"], years.loc[2022, "Job Stayer"]):
    print(f"   {g:.1f}% a year for five years = {((1 + g / 100) ** 5 - 1):.0%} in total")

W1 median wage growth of people who stayed in the same job (3-month moving average), yearly mean
   2019  stayers 3.2% | switchers 4.3%
   2020  stayers 3.2% | switchers 4.2%
   2021  stayers 3.4% | switchers 4.5%
   2022  stayers 5.5% | switchers 7.5%
   2023  stayers 5.3% | switchers 6.3%
   2024  stayers 4.7% | switchers 5.0%
   2025  stayers 4.1% | switchers 4.2%
   2026  stayers 3.5% | switchers 4.5%

W2 professional and management occupations, 12-month moving average to Sep 2026: 3.9% (all workers 3.7%)

W3 what a stayer's typical raise would mean: the same pay growth sustained for five years
   3.5% a year for five years = 19% in total
   5.5% a year for five years = 31% in total

For people who stayed in their jobs, median pay growth was about 3.2% a year in 2019 and 2020, rose to 5.5% in 2022 when employers were competing hard for staff, and was 3.5% on average in 2026 so far. Professional and management occupations were slightly above all workers, at 3.9% in the year to September 2026.

Two things follow. First, a typical increase for someone who stays is modest; a raise well above it needs a reason, such as a bigger role, rare skills or pay that has fallen behind the market. Second, small differences compound: 3.5% a year for five years adds up to 19%, while 5.5% adds up to 31%. A well-argued request that lifts your increase by a couple of points matters over a career. This is US data; Indian increments follow their own cycles and norms, but the logic of compounding is the same.

Timing: ask before the budget is set

GitLab's public handbook describes a typical model: an annual compensation review in which managers reflect on accomplishments and measure achievements against goals, with results shaped by individual performance, company performance, the available budget and local pay practices (GitLab: compensation review).

If your company works like this, asking during the review meeting is too late; the decision has been made. Better moments:

• Six to eight weeks before the review cycle, when your manager is preparing proposals.
• Right after a clear win, such as a project shipped or a result your stakeholders noticed.
• When your role has grown, for example you now own a system, lead others or cover a second area.

Avoid asking during company-wide cost cutting or immediately after a difficult project. Find out when your company's cycle runs; your manager or HR can tell you.

Prepare the evidence

Build a one-page summary. Your manager will often need to forward the case to someone else, so make it easy:

1. Results: three to five achievements with measurable impact: time saved, revenue or cost affected, decisions changed, incidents prevented.
2. Scope: how your responsibilities have grown since your last pay change.
3. Market context: what similar roles pay, from several sources, such as recent offers people you trust have received, salary surveys and job postings that list pay. Use ranges, not a single number from one website.
4. Next year: what you plan to take on.

Your evidence log makes this much easier.

A script

"I'd like to talk about my compensation before this year's review. Over the past year I've taken on the forecasting pipeline and the weekly business review, and two results stand out: the pipeline cut the monthly close by two days, and the churn analysis led to the new retention offer. Based on that scope and what similar roles pay, I'd like to ask for an increase above the standard adjustment. What would you need from me to support that?"

The last question matters. It turns a demand into a joint plan, and if the answer is no, you learn what would change it.

Mistakes that weaken the request

The most common mistake is arguing from need rather than value: rent, a family event or a friend's salary are real pressures, but they give your manager nothing to take to a budget meeting. The second is surprise: a manager who first hears about your request in the review itself has no room to act. The third is an ultimatum you do not mean. Threatening to leave can work once, but only if you are prepared to go, and it changes the relationship either way. Finally, avoid comparing yourself by name with colleagues; talk about your own scope, results and the market instead.

If the answer is no

• Ask why, and what would make it a yes, and agree a date to review.
• Ask about other levers: a title change, a bonus, training budget, or a larger project that sets up next year's case.
• Decide your own next step. If the gap with the market is large and nothing changes, it may be time to look elsewhere; when to stay and when to switch helps with that decision.

For negotiating a new offer rather than a raise, see salary negotiation.

Skills that strengthen your next pay conversation are part of the Data Analytics course.

Sources

• Wage Growth Tracker (Federal Reserve Bank of Atlanta) — Federal Reserve Bank of Atlanta
• Annual Compensation Review Cycle (The GitLab Handbook) — GitLab