How to Research What You Should Actually Be Paid

Employers have historically had a significant information advantage in salary negotiations: they know what every comparable person in the company earns, what they have paid previous hires, and what the budget for the role is. Candidates, by contrast, typically have limited data — what they currently earn, what a friend told them, or a generic range from a salary survey.

This information asymmetry consistently produces outcomes that favor employers. Candidates who do not know their market rate either undersell themselves to avoid the risk of asking too much, or make uninformed guesses that may miss the actual market by tens of thousands of dollars in either direction. Salary transparency laws in many states and countries are gradually closing this gap, but the best defense is doing your own research before any compensation conversation.

Levels.fyi. For technology roles specifically, Levels.fyi is the most granular and reliable source of compensation data. It includes self-reported total compensation broken down by company, role, level, and location — including base salary, bonus, and equity. If you are a software engineer, PM, or data professional, this is the first place to look.

Glassdoor and LinkedIn Salary. More general than Levels.fyi but broader in industry coverage. Glassdoor salary reports are self-reported by employees and tend to skew toward base salary. LinkedIn Salary provides similar data with a slightly different methodology. Use both and triangulate between them.

Bureau of Labor Statistics Occupational Outlook Handbook.The BLS provides government-collected compensation data for hundreds of occupations, broken down by industry and geography. It is less granular than crowdsourced sources but is based on systematic data collection rather than self-reporting, which means it catches some biases that self-reported data misses.

Salary transparency in job postings. An increasing number of jurisdictions now require employers to post salary ranges. Read these carefully — the ranges are often wide, but seeing what a specific company budgets for a specific level is more reliable than any survey data.

Raw compensation data requires interpretation. A salary figure for a “software engineer” in San Francisco tells you very little about what a mid-level backend engineer with five years of experience and specific expertise in distributed systems should expect at a Series C startup in Austin. The generic number needs to be calibrated for level, location, company size, and industry.

Location adjustments are significant. Compensation in major tech hubs (San Francisco, Seattle, New York) is typically 20-50% higher than equivalent roles in most other US markets. Remote roles at companies based in high-cost markets vary widely — some pay location-adjusted rates, others pay San Francisco rates regardless of where you live.

Company size and stage matter too. Large public tech companies pay at or above the top of the market range with substantial equity in the form of liquid RSUs. Early-stage startups often pay below-market base in exchange for larger equity grants. Financial services firms have different compensation structures from tech companies even for identical technical roles. Make sure your market data comes from companies comparable to the one you are evaluating.

Public compensation data is valuable but incomplete. The most granular and timely data comes from direct conversations with people doing comparable work — peers at similar companies in similar roles. These conversations require trust and reciprocity: you share your data, they share theirs.

Many people feel uncomfortable discussing compensation directly, but the discomfort is slowly eroding as transparency norms shift. Opening the conversation with your own number removes most of the awkwardness: “I make $X in this role — do you feel comfortable sharing your number so I can calibrate?” Most people who would be unwilling to volunteer their salary will share it when asked directly in a context of reciprocal sharing.

Anonymous forums and Slack communities for specific professional communities (Women in Product, particular engineering communities, industry-specific groups) often have channels dedicated to compensation sharing. These are particularly valuable because they tend to have high participation and real, specific numbers from identifiable professional contexts.