A pay conversation without research is two people guessing, and the side with more information usually wins. The information gap has narrowed enormously: what roles pay used to be a closely held secret, and now a lot of it is searchable in an afternoon. Here is how to do that homework, turning scattered numbers into a defensible range, so that any future conversation rests on evidence instead of nerves.
Where the numbers live
No single source is gospel, so the move is to triangulate: gather several numbers and look for where they cluster. Each source has a bias worth knowing.
| Source | What it tells you | Watch out for |
|---|---|---|
| Salary database sites | Crowd-sourced ranges by title and city | Self-reported, sometimes stale or inflated |
| Posted job listings | Real, current bands (required in a growing number of states) | A wide range can hide the realistic midpoint |
| Leveling guides | How titles map across companies | Titles don't translate cleanly between firms |
| People in the field | The most specific picture | Small sample; norms vary on whether to ask |
| Government wage data (BLS) | Broad, reliable medians by occupation | Lags the market; not company-specific |
The goal isn't one perfect number. It's a cloud of numbers from different angles. When a database, a few job postings and a knowledgeable contact all point at roughly the same zone, that zone is the market rate. When they scatter wildly, that itself is information: the role may be defined differently across companies, or leveled inconsistently.
Location, remote work and cost of living
The same role rarely pays the same everywhere. Companies set pay against local markets, so a job in a high-cost metro typically carries a higher number than the identical job in a lower-cost area. With remote work, employers handle this differently: some pay one national rate, others adjust by where you live.
Cost of living enters here, carefully. A higher number in an expensive city isn't automatically a better deal once rent and everyday costs are subtracted, and a lower number in a cheaper area can stretch further. Comparing two offers in different places means comparing what's left after the basics, not the headline figure. Inflation matters too: a number that looked strong three years ago has lost ground if pay didn't keep pace with prices.
Total comp: the number that actually matters
The single biggest mistake in pay research is comparing base salary alone. The real figure is total compensation: everything of value the job provides in a year.
| Component | What it is | How real the dollars are |
|---|---|---|
| Base salary | The guaranteed gross income on the paycheck | Fully real, predictable |
| Bonus | A variable, often performance-based add-on | Real but not guaranteed |
| Equity / stock | An ownership stake, equity that may vest over years | Real only if it vests and holds value |
| Retirement match | Employer match on a 401(k) | Free money, but only after vesting |
| Benefits and PTO | Health coverage, paid time off, perks | Real value, easy to undercount |
A job with a lower base but a strong employer match, good health coverage and more paid time off can out-value a higher-base job with thin benefits. The reverse is also true: equity that never vests or a bonus that rarely pays out makes a flashy "total" number far softer than it looks. The lesson on your benefits, 401(k) and insurance covers what those benefit lines are worth.
What you need versus what the market pays
One more line is worth drawing sharply, because blurring it weakens every future conversation. "What I need to cover my life" and "what this role pays in the market" are different questions. Your household budget tells you the floor you can't fall below, but the market doesn't pay based on your rent. It pays based on the value and scarcity of the work.
Anchoring a pay expectation to personal expenses tends to leave money on the table, because the market figure is often higher than the "I just need to get by" number. The two are useful for different things: the need number is a safety floor (the budget calculator maps it), and the market number is what an evidence-based target is built on. Keeping them separate is what makes a range defensible rather than emotional. The next lesson turns that researched range into an actual conversation.