Compensation Benchmarking

The process employers use to compare their pay levels against market data — and the same process employees can use to verify their compensation is competitive before negotiations.

Compensation benchmarking is the process of systematically comparing a company's pay levels for specific jobs against external market data to determine whether compensation is competitive. Employers use benchmarking to set salary ranges, make offer decisions, and conduct annual pay reviews. The process relies on compensation survey data — structured surveys collected by firms like Radford (Aon), Willis Towers Watson, Mercer, and the Economic Research Institute — that aggregate actual pay data from large numbers of participating companies, broken down by job function, level, industry, company size, and geography. Participation in these surveys typically requires a company to share their own compensation data in exchange for access to aggregate market data.

The benchmark process starts with job matching — aligning internal job descriptions to standardized survey job definitions. A 'Software Engineer III' at one company must be matched to the appropriate benchmark job (Radford calls it 'Software Engineer - Level 3') before the market data is meaningful. Job matching is an art as much as a science: job titles vary enormously across companies, and the actual scope of a role (people management, budget responsibility, strategic influence) determines the right benchmark match, not just the title. Mismatched benchmarking produces misleading results in either direction — a senior engineer benchmarked to a junior role understates market pay; a mid-level manager benchmarked to a director role overstates it.

Once benchmarks are established, companies compare their pay to market percentiles — typically targeting the 50th, 75th, or 90th percentile depending on their compensation philosophy. A company targeting the 75th percentile intends to pay above 75% of the market for each role. Compa-ratio is the key metric: your actual salary divided by the market midpoint (50th percentile). A compa-ratio of 1.0 means you're paid exactly at market median; 0.85 means you're 15% below market; 1.15 means you're 15% above. HR teams monitor compa-ratios across the organization to identify who is paid significantly below or above market and calibrate merit increases accordingly.

Employees can do their own benchmarking — imperfectly, but meaningfully — using public data sources. Levels.fyi provides verified compensation data for tech roles at specific companies, including base, bonus, and equity details. LinkedIn Salary provides aggregate ranges by title and location. Glassdoor provides self-reported compensation (lower reliability but broader coverage). The Bureau of Labor Statistics Occupational Employment Statistics program provides aggregate national and regional data by occupation. The most effective employee benchmarking uses multiple sources, focuses on roles with similar scope and level, controls for geography and company stage, and focuses on total compensation rather than just base salary — since equity and bonus represent large fractions of total pay in many industries.

How to Benchmark Your Own Pay

  • Use multiple sources: Levels.fyi (tech, verified), LinkedIn Salary (broad, aggregate), Glassdoor (self-reported, broad), Radford/WTW published summaries, BLS OES data.
  • Control for level, not just title: a 'Senior Engineer' at a 10-person startup and a 'Senior Engineer' at Google are not the same job — level calibration (L5, L6, etc.) matters more than title.
  • Geography matters significantly: a software engineer role in San Francisco pays 40-60% more than the same role in Austin or Denver — always filter by location when benchmarking.
  • Total compensation: compare total comp (base + target bonus + annualized equity value), not just base salary — in many tech roles, equity and bonus represent 30-60% of total comp.
  • Company stage: early-stage startups typically pay below market on cash but compensate with equity and upside; mid-stage companies (Series B-D) compete more aggressively on cash; public companies compete fully on all dimensions.
  • Benchmark to the 50th and 75th percentile: knowing where you fall against both midpoint (50th) and competitive (75th) gives you a range to reference in negotiations.

How Companies Use Benchmarking

  • Setting salary ranges: benchmark data defines the midpoint of salary ranges; the range typically extends ±20% around the midpoint (80th to 120th percentile of the midpoint).
  • Offer decisions: recruiters compare offer amounts to the benchmark for the level; offers significantly above range require additional approval.
  • Merit increase calibration: employees below 90% of the benchmark midpoint (compa-ratio < 0.90) may receive above-average merit increases; those above 110% may receive below-average increases.
  • Pay equity audits: companies use benchmarking data to identify systematic gaps between demographic groups within the same job and level.
  • Annual market adjustments: survey data is updated annually; companies may give market adjustments (separate from merit) to bring pay in line with significant market movements.
  • The lag problem: survey data reflects compensation from 6-18 months ago; in fast-moving markets (tech salaries in 2021-2022), companies using lagged data fell behind market quickly.

Using Benchmarking in Negotiations

  • Cite specific data sources: 'According to Levels.fyi, the median total comp for an L5 engineer in Seattle is $X' is more persuasive than 'I've seen higher offers.'
  • Ask where the offer falls in the band: 'Can you tell me where this offer falls within the salary range for this level?' is a reasonable question; some companies will answer.
  • Focus on total comp: if base isn't moving, ask about signing bonus, equity, or performance bonus to improve total comp without changing the base salary line.
  • Use competing offers: a documented competing offer is the most effective benchmarking tool — it provides concrete, real-time market evidence that is hard to dispute.
  • Understand their philosophy: a company that targets the 50th percentile has different latitude than one targeting the 75th — the former has more room if you can make the case.

Example

A product manager with 7 years of experience receives a job offer for $155,000 base salary. She benchmarks: LinkedIn Salary shows a median of $168,000 for Senior PM roles in her city. Glassdoor shows a range of $140,000–$195,000. She also has a contact at the company who shared that their compensation targets the 60th percentile. She responds: 'I'm very excited about this role. Based on my research on LinkedIn Salary and a few recent offers I'm aware of in this market, I was expecting something closer to $170,000. Can we get there?' The recruiter comes back at $165,000 — a $10,000 improvement over the initial offer, with no competing offer leverage, purely from a prepared benchmarking conversation.