Benchmarking Your Pay

You cannot negotiate effectively if you don't know what you're worth. Most engineers have a vague sense that they're "probably paid okay" but have never actually checked. The data is available. The tools exist. There is no reason to operate on guesswork when you can know, with reasonable precision, whether you're being paid fairly.

Where to Find Compensation Data

Several platforms aggregate real compensation data. Each has strengths and weaknesses.

Source            Strengths                       Weaknesses
levels.fyi        Verified offers, level-specific  Skews toward big tech
Glassdoor         Broad coverage, many companies   Self-reported, often outdated
Blind/Teamblind   Real-time, anonymous             Skews high, selection bias
Payscale          Statistical models               Less tech-specific
LinkedIn Salary   Large dataset                    Broad ranges, less useful
Salary.com        Enterprise data                  Skews corporate/non-tech
H1B Salary DB     Actual filed salaries            Only H1B holders, base only

How to Use levels.fyi

levels.fyi is the gold standard for tech compensation benchmarking. Here's how to use it effectively:

Step 1: Find your company and level
        Google L5, Meta E5, Amazon L6, etc.

Step 2: Look at total compensation, not just base
        TC = Base + Equity + Bonus

Step 3: Check the range, not just the median
        Median: $350k
        25th percentile: $310k
        75th percentile: $400k
        Your position in this range tells you something.

Step 4: Filter by location and years of experience
        A Bay Area L5 with 8 years experience should be
        compared to Bay Area L5s with similar experience,
        not to everyone globally.

Step 5: Check the date
        Comp data from 2021 (peak market) is different
        from 2023 (post-correction). Use recent data.

Blind is an anonymous professional network where tech workers share compensation data freely. It is unfiltered, which makes it both valuable and noisy.

What Blind is good for:
  - Real-time market data during active hiring seasons
  - Understanding negotiation outcomes ("I got X by doing Y")
  - Company-specific compensation culture
  - Learning about refresher grant norms
  - Understanding what competing offers look like

What to watch out for:
  - Selection bias: people who post tend to have higher comp
  - Bragging: some posts are more flex than data
  - Narrow demographic: heavily skews Bay Area, big tech
  - Toxicity: filter the noise, extract the signal

H1B Salary Database

The H1B salary database contains actual salaries filed with the Department of Labor. These are real numbers, not self-reported. The limitation is that they only include base salary (no equity, no bonus) and only cover H1B visa holders.

Search: h1bdata.info or h1bsalarydata.com

Example search: "Software Engineer" at "Google" in "Mountain View"

Results show:
  Job title, base salary, location, year

Useful for: establishing a floor for base salary at specific companies
Not useful for: total compensation, equity data, non-visa holders

How to Know If You're Underpaid

Collect data points from multiple sources and compare them to your current compensation:

Your current situation:
  Company: Mid-stage startup
  Level: Senior Software Engineer
  Location: San Francisco
  Years of experience: 6
  Base: $175,000
  Equity: $30,000/year (paper value, pre-IPO)
  Bonus: None
  TC: $205,000 (cash: $175,000)

Market data (levels.fyi, Blind, peers):
  Senior SWE, SF, 6 YOE at public companies:
  Median TC: $340,000
  25th percentile TC: $290,000
  Cash comp (base + bonus): $210,000 - $250,000

Assessment: Underpaid by $85-135k in TC.
            Even comparing cash-only, below 25th percentile.
            The equity is speculative and illiquid.

Signs you're likely underpaid:

You haven't changed jobs in 3+ years and your raises have been 3-5% annually
You were hired when you had less experience and your level hasn't changed
Peers with similar experience who joined recently earn noticeably more
You've never negotiated your salary
Your company doesn't do market adjustments

When to Ask for a Raise vs When to Leave

This is a decision tree, not a feeling:

Is your company able to pay market rate?
  |
  +-- No (small startup, non-tech company, budget constraints)
  |   -> Leave. No amount of asking will create budget that doesn't exist.
  |
  +-- Yes
      |
      Is your manager supportive and influential?
        |
        +-- No (manager is unhelpful or powerless)
        |   -> Leave, or find a new team internally first.
        |
        +-- Yes
            |
            Have you been given a clear path to higher comp?
              |
              +-- No -> Ask for a raise with market data. Give it one cycle.
              |         If nothing changes, leave.
              |
              +-- Yes -> Execute the path. If the raise materializes,
                         stay. If it was empty promises, leave.

The Internal Raise Process

Asking for a raise internally requires preparation:

What you need:
  1. Market data (levels.fyi, competing offers)
  2. Your performance evidence (shipped projects, impact metrics)
  3. Your level and band position (ask HR or your manager)
  4. Timing (during comp review cycle, not randomly)

What to say:
  "Based on market data from [sources], the median TC for a 
  [level] engineer with my experience in [location] is [$X]. 
  My current TC is [$Y]. I'd like to discuss bringing my 
  compensation to market rate."

What NOT to say:
  "My friend at Google makes more than me."
  "I need more money because my rent went up."
  "I'll leave if you don't pay me more." (unless you mean it)

Most companies have a compensation review cycle once or twice a year (typically Q1 or Q4). Making your case outside of this cycle is harder because the budget is already allocated. Time your ask to land 2-4 weeks before the review cycle starts, so your manager can advocate for you when decisions are being made.

Market Timing

The tech job market has cycles. Your ability to get a raise or a better offer depends partly on market conditions:

Strong market (2020-2021 style):
  - Companies competing aggressively for talent
  - Counter-offers are generous
  - Signing bonuses are inflated
  - Equity grants are large
  - Good time to negotiate or move

Weak market (2023 style):
  - Layoffs and hiring freezes
  - Fewer competing offers available
  - Companies less willing to match
  - Negotiation leverage reduced
  - Better to stay unless actively at risk

Neutral market:
  - Normal hiring pace
  - Standard negotiation dynamics
  - Good time to move if underpaid

Market conditions don't change the fundamentals, but they affect your leverage. In a weak market, you might need to accept a smaller improvement and wait. In a strong market, the opportunity cost of not shopping around is enormous.

The Cost of Loyalty

Staying at one company for a long time is often portrayed as a virtue. In terms of compensation, it is usually a mistake.

Engineer who stays 6 years at one company:
  Year 1: $250,000 TC
  Year 2: $260,000 (+4% raise)
  Year 3: $270,000 (+4% raise)
  Year 4: $280,000 (+4% raise)
  Year 5: $290,000 (+4% raise)
  Year 6: $300,000 (+4% raise)
  6-year total: $1,650,000

Engineer who switches every 2 years:
  Year 1-2: $250,000 TC ($500,000)
  Year 3-4: $310,000 TC ($620,000) <- 24% jump from new offer
  Year 5-6: $370,000 TC ($740,000) <- 19% jump from new offer
  6-year total: $1,860,000

Difference: $210,000 over 6 years.
Plus: two signing bonuses, two fresh equity grants.

The job-hopper earned $210k more over the same period by simply accepting market-rate offers every two years. This doesn't account for signing bonuses or the higher base they're building on going forward.

When Loyalty Does Pay

Staying can be rational in specific situations:

You're on a strong equity trajectory (early at a company heading toward IPO)
You're close to a promotion that would be hard to get externally
You have unusual benefits (fully remote, exceptional team, rare work-life balance)
You value stability and the premium for leaving isn't large enough
You're vesting a large equity package and refreshers are strong

The key is to make it a conscious choice based on data, not inertia.

How to Benchmark Regularly

Make benchmarking a recurring process, not a one-time event:

Annually (during comp review season):
  - Check levels.fyi for your level and location
  - Review Blind for recent data points
  - Talk to trusted peers about market rates (be willing to share yours)
  - Compare your TC to the 50th percentile

Every 2-3 years:
  - Do a real market test: interview and get offers
  - Even if you don't plan to leave, knowing your market value
    is information you need
  - A competing offer is the strongest leverage you can have

When changing jobs:
  - Research the specific company's comp bands
  - Get multiple offers when possible
  - Use levels.fyi offer data to calibrate expectations

Common Pitfalls

Only comparing base salary. TC is the metric. Two jobs with identical base salaries can differ by $100k+ when equity and bonuses are included.
Using stale data. A data point from 2021 is not relevant in 2025. Markets shift. Use data from the last 6-12 months.
Selection bias in your reference group. If all your friends work at FAANG and you work at a Series B startup, their comp is not your benchmark. Compare like for like: same level, same type of company, same location.
Never actually testing the market. You can look at data all day, but the only way to know your true market value is to get a real offer. Data tells you the range. An offer tells you where you land.
Staying out of inertia. "I'm comfortable here" is not a financial strategy. Comfort has a price. Know what that price is before deciding to pay it.
Thinking you can't leave because of unvested equity. Calculate the cost of staying (below-market pay) vs the cost of leaving (unvested equity). Often, a new offer more than compensates for what you leave behind.

Key Takeaways

Use levels.fyi, Blind, and H1B salary data to benchmark your compensation. Cross-reference multiple sources for accuracy.
If you haven't changed jobs in 3+ years and haven't received a significant market adjustment, you are probably underpaid.
Internal raises (3-5%/year) almost never keep pace with market movement. Job changes typically yield 15-25% increases.
Time your raise requests to coincide with the company's compensation review cycle. Bring market data, not emotional arguments.
The cost of loyalty is real and measurable. Staying at one company for 6+ years without market adjustments typically costs $100-300k in cumulative compensation versus strategic job changes.