The Number Most People Leave on the Table
Meet Priya. She's a software engineer, 4 years at the same company, currently earning $92,000. She walks into her review and asks for "a raise." Her manager offers 3%. She takes it. That's $2,760. She feels good for about a week.
Here's the math she never ran. The market rate for her role and experience in her city is $108,000. She is underpaid by $16,000 — not because she asked and was refused, but because she asked for a feeling instead of a figure. The number she should have brought into the room was a 17% adjustment to market, not a vague "more."
This is the trap nobody warns you about: companies budget raises as a percentage of your current salary, not your market value. If you started underpaid, every annual 3% bump compounds the gap. After 4 years, Priya's 3%-a-year raises kept her permanently below the rate a new hire would command for her exact job.
What a defensible raise number actually looks like. A good ask has three components stacked on top of each other:
- The market gap — the difference between your salary and the median for your role, level, and location. This is the foundation. For Priya, $16,000.
- Performance premium — if you've shipped measurable results, you justify the top of the band, not the median. A documented win like "cut deployment time 40%" or "closed $400K in new accounts" moves you 5–10% above median.
- Cost-of-living and inflation — with the 2026 cost of living climbing, a raise that just matches inflation isn't a raise at all. Your purchasing power has to grow, not tread water.
Stack those and Priya's defensible ask isn't 3%. It's a move from $92,000 to roughly $112,000 — market median plus a performance premium for her shipped work. Even if she anchors high and lands in the middle, she captures $14,000 she would otherwise have walked past.
Why the specific number wins. "I'd like a raise" invites a counter-offer of whatever fits the budget. "Based on market data for senior engineers in this metro, the rate is $108,000, and my work this year supports the top of that band" invites a negotiation about a number you chose. You set the anchor. That's the entire difference between a 3% bump and a 17% correction.