Nothing makes a salary conversation scarier than not knowing the number. Name a figure too low and you can leave thousands on the table for years. Name one too high with nothing to back it up and you look out of touch. The good news: the fix is not confidence or a clever script. It is research, and it is far more doable than most people think.
Know what "market salary" actually means
Before you learn how to research a salary range, get clear on what you are looking for. Market rate is not a single number. It is a range shaped by five things: your role and level, your location, your industry, the company's size and funding stage, and your specific skills. Two people with the same title can be paid very differently and both can be "at market."
So the goal is not to find one magic figure. The goal is to build a defensible range for your exact situation. When you can point to where each number came from, you stop guessing and start negotiating from evidence.
Start with free government data
Your most credible and least biased baseline is public data. In the US, the Bureau of Labor Statistics runs the Occupational Employment and Wage Statistics (OEWS) program, free at bls.gov. Search your occupation and you get the median wage plus the 10th, 25th, 75th, and 90th percentiles, both nationally and for specific metro areas.
That percentile spread is the useful part. If the 25th to 75th percentile for your role in your city runs from $72,000 to $96,000, you now have an honest anchor. The limitations: BLS uses broad occupation categories, the data lags a year or two, and it reports base wages only, not bonus or equity. Treat it as your floor-setting reference, not the whole picture.
Pull real numbers from current job postings
This is the step most people skip, and it is the strongest signal you can get. A growing number of US states and cities now require employers to list pay ranges in job postings. That means you can read the exact ranges companies are advertising right now for the exact role you want.
Search open listings for your target title in your area and collect eight to twelve postings. Write down each posted range. Watch for level mismatches, because a "Marketing Manager" at a 30-person startup and at a Fortune 500 can be different jobs. Look at where the ranges cluster and where the midpoints land. This is the freshest, most role-specific data available, sourced straight from employers, so weight it heavily.
Cross-check with salary aggregators
Aggregator sites help you confirm your range, not set it. Use several and read each one critically:
- Levels.fyi is the best source for tech total compensation, including base, bonus, and equity by company and level.
- Payscale and Salary.com generate personalized estimates from a short questionnaire.
- Glassdoor, LinkedIn Salary, Comparably, and Indeed rely on self-reported figures, so treat them as directional.
The honest caveat: self-reported data suffers from selection bias, stale entries, and title inflation. Any single site can be off by a lot. When three independent sources agree on a range, you can trust the overlap. When one is an outlier, set it aside.
Adjust for the factors that move pay
Raw numbers still need tuning to your reality:
- Location and remote policy. Pay tracks the local cost of labor. A fully remote role may be benchmarked to a national average or to the company's headquarters market, so ask which.
- Level, not title. Map the actual responsibilities, not the label. A "senior" title at one company equals a "mid-level" scope at another.
- Industry. The same skills often pay more in finance or software than in nonprofits or education.
- Company size and stage. Early startups may offer lower base and more equity. Large public companies usually pay steadier cash.
- Your leverage. In-demand skills, a specialized certification, or a competing offer all push you toward the top of the range.
Also check whether your profession has an annual salary survey. Many professional associations publish one, and it is often more precise for a niche role than any general site.
Count total compensation, not just base
Base salary is only part of the offer. To compare roles fairly, add up total compensation: base pay, annual bonus, equity or RSUs, 401(k) match, health benefits, PTO, and any sign-on bonus. A job with a lower base but a real bonus and a 6 percent match can beat a higher base with none of that. Convert everything to a comparable annual number so you are comparing like with like.
Build your range: floor, target, and stretch
Now turn the research into three numbers:
- Floor: your walk-away point, grounded in the lower percentiles and the bottom of posted ranges.
- Target: the realistic ask your data supports, near the median for your level and location.
- Stretch: the top of market you would gladly accept, backed by your leverage.
Here is how that looks in practice. Maya is a Customer Success Manager in Denver with five years of experience. BLS puts the Denver-area median for her role near $90,000, with the 75th percentile around $108,000. She collects ten local job postings, which advertise ranges from $85,000 to $112,000. Levels.fyi and Payscale both land her in the mid-$90,000s. She sets her floor at $88,000, her target at $98,000, and her stretch at $110,000. When a recruiter asks about pay, she can name $95,000 to $105,000 and explain exactly where it comes from.
Bring your research to the table
A well-researched number lands best when your experience is presented at the right level in the first place. If your resume undersells your scope, you get benchmarked to a lower band before you ever say a number. It is worth making sure your resume shows the seniority you actually carry, in language a hiring system reads cleanly. You can see the literal text an applicant tracking system pulls from your resume with the free Beat the Bots scan at careerbounce.io, so nothing about your level gets lost in parsing.
And when you build the resume itself, honesty is your ally in the negotiation that follows. Bounce Studio writes an ATS-ready resume using only your real experience, verified never to invent skills or tools you do not have. That matters here for one reason: the strongest position in any salary conversation is a story you can fully defend in the interview. Do the research, know your range, and walk in with numbers and a resume you can both stand behind. Everyone bounces back, and the ones who negotiate from evidence tend to bounce back further.
FAQ
How many sources do I need to research a salary range? Aim for at least three independent sources: government data like BLS, current job postings with posted ranges, and one or two aggregator sites. When they overlap, you have a reliable range. If one number is far off from the rest, treat it as an outlier rather than proof.
Is Glassdoor accurate for salary research? Glassdoor is useful as a directional signal, not a final answer. Its figures are self-reported, so they can skew high or low and go stale. Use it to confirm a range you built from stronger sources like BLS and live job postings, not to set the number on its own.
How do I research salary for a fully remote job? First ask, or check the posting, whether the company benchmarks remote pay to a national average or to its headquarters market, because that changes the range a lot. Then pull posted ranges from remote listings for the same role and level, and compare them against BLS data for the relevant metro. Total compensation still matters, so factor in bonus and benefits.
What if I cannot find data for my exact job title? Titles vary, so match on responsibilities and level instead of the label. Find the closest standard occupation in BLS, then triangulate with job postings that describe similar duties. Professional association salary surveys are also helpful for niche or specialized roles.
Should I share my salary research during the negotiation? You do not need to hand over your spreadsheet, but naming a range and briefly citing where it comes from makes your ask credible. Saying your target reflects market data for your role, level, and location signals that you have done the work. That framing is harder to dismiss than a number with no basis behind it.