Benchmarking salaries in the United States
Reviewed by Mellow Editorial Team, HR & payroll content team
Benchmarking a salary means checking what the market actually pays for a role, so you can set pay that attracts and keeps good people without overpaying or underpaying. Done well, it gives you a defensible number backed by data rather than instinct.
Why benchmarking matters
Pay that drifts below market leads to turnover. Pay that runs well above it strains your budget without necessarily improving retention. Neither outcome is good.
Benchmarking also helps you avoid pay equity problems. If you can show that compensation decisions follow a consistent methodology tied to market data, you have a stronger defense against discrimination claims — something worth considering given the growing number of pay transparency laws at the state and city level.
Where to find reliable salary data
No single source is perfect, so most HR leads combine two or three.
Government data. The Bureau of Labor Statistics publishes the Occupational Employment and Wage Statistics (OEWS) survey — free, comprehensive, and updated annually. It breaks down median and percentile wages by occupation, industry, and metropolitan area. It lags the market by six to twelve months, but it is authoritative and defensible.
Commercial surveys. Radford (now part of Aon), Mercer, Willis Towers Watson and the Economic Research Institute all publish detailed compensation surveys. These cost money to access, but they are updated more frequently and slice the data by company size, funding stage and industry vertical — useful if you are a startup competing against large employers for the same talent.
Free aggregators. LinkedIn Salary, Glassdoor, Levels.fyi (for tech roles) and Payscale give you directional data quickly. Treat these as a sanity check rather than a primary source — self-reported data has obvious biases.
Peer networks. Founders and HR leads in industry groups often share anonymized compensation data informally. This is especially useful for early-stage companies where formal survey participation is impractical.
How to define the right comparison set
A benchmark is only useful if you are comparing yourself to the right employers. Define your comparison set before you pull any numbers.
Industry. A software engineer at a fintech startup commands different pay than a software engineer at a regional manufacturer. Use industry filters whenever the data source allows.
Geography. Labor markets vary enormously across the US. A marketing manager role in San Francisco or New York City will benchmark significantly higher than the same role in Tulsa or Memphis. If you have remote employees, decide in advance whether you pay to the employee's location, your headquarter's location, or a hybrid — and apply that policy consistently.
Company size and stage. Compensation at a 20-person Series A startup looks different from compensation at a 5,000-person public company. Smaller companies often compensate for lower cash salaries with equity; factor that in when making comparisons.
Job level. Titles are notoriously inconsistent across companies. A "Senior Engineer" at one firm might be a "Staff Engineer" at another. Match on the actual scope of the role — number of direct reports, budget responsibility, technical complexity — not just the title.
Turning data into a pay structure
Once you have market data, the standard approach is to build salary bands — a minimum, midpoint and maximum for each role level.
The midpoint is typically set at the 50th percentile (median) of your target market, though companies competing aggressively for talent sometimes set it at the 75th percentile. The band then stretches a set percentage above and below that midpoint — commonly 20–25% in either direction, though this varies by industry and role type.
Within a band, individual placement depends on factors like tenure, performance, and how quickly someone is developing in the role. Document these criteria. Ad hoc decisions made without a framework are the fastest route to pay inequity and legal exposure.
Revisit your bands at least once a year. Market rates for certain roles — particularly in technology, data science and AI — can shift materially in twelve months.
Accounting for total compensation, not just base salary
Base salary is the most visible part of an offer, but the full picture includes employer contributions to health insurance premiums, retirement plan matching, equity or profit-sharing, and any bonuses. When you benchmark, make sure you know whether the survey data you are using reflects base salary only or total cash compensation — mixing the two produces misleading comparisons.
Also keep payroll costs in mind. Employers pay their share of FICA — matching the employee's 6.2% Social Security contribution (up to the annual wage base) and 1.45% Medicare contribution — on top of every dollar of salary. That adds roughly 7.65% to your base payroll cost before you factor in benefits, workers' compensation insurance or any state-specific taxes. A fully-loaded cost model gives you a clearer picture of what a hire actually costs the business.
For companies hiring across multiple states or countries, how Mellow runs payroll across six countries on one platform gives a practical sense of how to manage that complexity without building separate processes for every jurisdiction.
Keeping your benchmarks current
Salary data has a shelf life. Set a calendar reminder to review your compensation bands annually — or sooner if you are losing candidates at the offer stage or seeing unusual attrition in specific roles. Both are strong signals that your market data has gone stale.
---
Run HR and payroll in United States with Mellow
Mellow brings HR, payroll and 12 AI agents into one platform — built to handle United States properly, with payroll included, from £4 per employee per month. The AI agents don't just answer questions; they generate contracts, run cost estimates and draft letters for you.
- United States payroll software
[Start a free trial →](/register)