Benchmarking your worth as a product manager means triangulating three imperfect sources—public aggregators, recruiter-grade surveys, and structured peer disclosures—then adjusting the blended number for your actual level, your company's stage, and your metro. No single source is trustworthy alone. The discipline lives in the cross-check, not the lookup.

Quick Answer: Pull a public range from levels.fyi or Blind, cross-check it against a recruiter-quoted or Radford-style survey range, then adjust for your company's stage (cash-heavy vs. equity-heavy) and metro tier. Triangulate at least three sources before you trust any single number enough to say it out loud.

Why Public Salary Data Lies to You (Even When It's Not Lying)

Public comp data isn't fabricated, but it's badly skewed by who bothers to report it: people who negotiated well, landed brand-name offers, or hit a title milestone self-select into posting, while quietly underpaid PMs stay invisible. Layer in stage-mixing—a Series B "Senior PM" and a Series G "Senior PM" pooled under one median—and a single number turns misleading.

Survivorship Bias: The Data You See Is the Data That Bragged

Think about who actually submits a data point to a public salary tool. It's disproportionately someone who just got a competing offer, someone who moved to a household-name tech company, or someone venting after a bad review cycle—not the median PM quietly doing solid work at a mid-size company for three years.

levels.fyi itself has been candid that its dataset skews toward large, well-known tech employers because submissions are voluntary and self-selected. Blind's anonymous forum skews toward people already primed to compare notes, often mid-negotiation. Neither is dishonest data—both are incomplete populations dressed up as representative ones.

  • Companies with strong brand recognition are overrepresented relative to their share of total PM headcount.
  • People currently negotiating (and therefore motivated to check numbers) are overrepresented relative to people who haven't looked in two years.
  • Non-tech-hub metros and smaller companies are underrepresented, which drags the visible median upward relative to the true median.

Stage-Mixing Bias: One Title, Five Different Jobs

"Senior Product Manager" at a 40-person Seed startup, a 400-person Series C, and a public company can be three different jobs wearing the same title—different scope, different equity risk, different cash-versus-stock mix. An aggregator that reports one blended number for "Senior PM" is averaging across all three, and the average describes none of them well.

This is the single biggest reason a number from a friend's offer, or a headline from a compensation aggregator, can feel wrong once you actually compare situations. It usually is wrong for your situation—not because the data point was invented, but because it was never meant to represent your stage.

Triangulating the Three Real Sources of Truth

No single source is built to give you a defensible number by itself; each one is strong on one axis and blind on another. The fix is to pull from all three—public aggregators, recruiter or survey-derived ranges, and structured peer conversations—and only trust the number where they converge.

Public Aggregators: levels.fyi, Blind, and LinkedIn Salary

These are your volume source: fast, free, and useful for a rough first anchor. levels.fyi is the most PM-specific and increasingly verifies submissions against offer letters for some companies, which helps. Blind adds real-time color from people currently interviewing, and LinkedIn's aggregate salary insights are broader but far less level-specific.

Use these to set a wide starting range, not a final number. Treat every figure as a self-reported outlier until proven typical.

Recruiter and Comp-Team Ranges: Radford, Carta, and Pave

This is your rigor source. Radford (an Aon compensation survey) is the backbone many corporate comp teams actually build internal bands from—PMs rarely see the raw survey, but recruiters quoting "our range for this level is X to Y" are usually paraphrasing something Radford-adjacent. Carta's Total Compensation data and Pave's benchmarking platform serve the same function for venture-backed startups, blending cash and equity by stage and headcount.

  • Ask recruiters directly for the leveled range, not just a single number—most comp teams have one and will share it once you're past the first screen.
  • Note what the range excludes. Sign-on bonuses, refresh grants, and target bonus percentages are frequently quoted separately from "base" or "TC."
  • Treat the quoted range as a starting anchor, not a ceiling—recruiter ranges are built to close a hire, not to maximize your outcome.

Peer Whispers: Structured, Not Gossip

This is your specificity source, and it's the one PMs use worst. A single overheard number from one peer is an anecdote; three to five peer conversations run with the same fixed questions become a pattern. Ask each person the same three things: base, target bonus percentage, and grant size relative to level—not a vague "so what do you make."

This is the same discipline as running structured discovery interviews the way our Jobs to Be Done guide describes: a handful of subjects, consistent questions, and a search for the pattern rather than the loudest anecdote. One data point is noise; five consistent ones are a signal you can act on.

Source typeBest forMain bias to watchHow to use it
Public aggregators (levels.fyi, Blind, LinkedIn)Fast, wide starting rangeSurvivorship — skews toward brand names and active negotiatorsSet the outer bounds, not the target
Recruiter / survey-derived (Radford, Carta, Pave)Rigor, real comp-team logicBuilt to close a hire, not maximize your numberAnchor the middle of your band
Structured peer conversationsSpecificity to your exact stage and metroSmall sample, one company's quirksConfirm or challenge the aggregator range

The table above is a lookup, not a scorecard—no source wins outright. A number that only one of the three sources supports is a hypothesis; a number all three roughly agree on is close to a fact.

Adjusting the Raw Number: Level, Stage, and Geography

A triangulated range is still wrong until you adjust it for three variables that move it more than any single data source will: your actual level, your company's stage, and your metro. Skipping this step is why two PMs with the same "Senior PM" title can have wildly different defensible numbers.

Level: Match Scope, Not Title

Titles inflate and deflate independently of scope across companies—a startup "Senior PM" might map to an L4 IC at a large tech company, or a "Group PM" at a scaling startup might really be doing L6-equivalent work. Comparing yourself to a title on an aggregator without first translating your actual scope into a shared framework is comparing apples to job descriptions.

Before you trust any benchmark, run your real scope—team size influenced, revenue or product surface owned, decision rights—through a consistent leveling lens. Our PM leveling rubric guide walks through how to do that translation so you're benchmarking the job you actually do, not the label on your badge.

Stage: Cash-Heavy vs. Equity-Heavy Bands

Company stage changes the composition of comp even when total value looks similar on paper. Early-stage companies pay more in speculative equity and less in cash; late-stage and public companies invert that ratio. Treating a Series A offer's headline "total comp" number as equivalent to a public company's is the fastest way to misprice an offer.

Company stageTypical cash weightingTypical equity weightingWhat to watch
Seed / Series ALower, often below-market baseHigh — mostly options, high risk/rewardStrike price, vesting cliff, dilution risk
Series B–CModerate — closer to market baseMeaningful — mix of options and RSUs at growth-stage valuationsLast-round valuation credibility
Series D+ / late growthNear or at market baseLower percentage, but larger absolute grantsWhether a liquidity event is realistically near
Public companyAt or above market base, bonus/OTE commonRSUs with real, tradable valueVesting schedule and refresh grant cadence

Directionally, this mirrors how compensation-survey providers like Radford and Aon report geo and stage differentials to their corporate clients—the underlying pattern (more cash certainty as risk drops) holds even though every individual company's exact mix varies.

Geography: Tiered Metro Adjustments

Most structured comp bands—whether from Radford-style surveys or a scale-up's internal comp tool—apply geographic differentials rather than one national number. Tier-one hubs (the Bay Area, New York, Seattle) sit at the top; secondary tech hubs (Austin, Denver, Boston) typically land a meaningful step below; and fully distributed or lower-cost metros land lower still, though the gap has narrowed since remote-first hiring became common.

  • Ask which model your target company uses: location-based bands, a single national band, or a hybrid with location-adjusted equity only.
  • Don't assume remote means national-tier pay. Many companies that pay "one national band" set it closer to the secondary-hub number, not the top-tier one.
  • Recheck this annually—differentials compress and expand as the broader hiring market shifts.

Worked Example: Building a Defensible Band for a Senior PM in Austin

Here's how the triangulation actually plays out for a concrete case: a Senior PM at a Series C fintech startup in Austin, benchmarking before a raise conversation. The process below is illustrative—swap in your own pulled figures—but the sequence of steps is what matters.

  1. Pull the aggregator range. A levels.fyi filter for Senior PM, Austin, growth-stage companies shows a wide spread—call it roughly $150K–$185K base, $185K–$235K total comp including equity value at last-round pricing.
  2. Cross-check against a recruiter or survey-derived range. A recruiter at a comparable Series C company quotes a leveled band of $160K–$180K base with a 10–15% target bonus, which narrows the aggregator's wider spread.
  3. Run three peer conversations. Two peers at similar-stage Austin startups report base salaries in the $165K–$175K range with option grants sized at roughly 0.05–0.08% of the company; a third, at an earlier-stage company, reports lower cash and higher percentage equity—consistent with the stage-mixing pattern above.
  4. Adjust for actual scope. Running the candidate's real responsibilities through a leveling rubric confirms they're doing genuinely senior-level work, not inflated-title mid-level work—so the band doesn't need a downward correction.
  5. Adjust for stage risk. Because Series C equity is still speculative, the candidate weights the band toward the cash figures and treats equity as upside rather than counted-on value.
PercentileBase salaryTarget bonusEquity (annualized, illustrative)Approx. total comp
25th$155K8%$15K~$182K
50th$168K10%$22K~$207K
75th$180K12%$30K~$232K

That table is the defensible band: a range, anchored in the 50th-percentile figure, with the underlying sources noted so it can survive a follow-up question. Walking in with "$207K, and here's how I got there" beats walking in with a single number pulled from one screenshot.

Turning a Band Into Negotiating Leverage

A triangulated band is only useful if you use the range, not a single point, and tie it to something a manager or recruiter can act on. Anchoring high within your defensible band—and being ready to explain the sources—works better than either lowballing yourself or quoting an unsupportable outlier number.

For an internal raise, the band is necessary but not sufficient; it needs to be paired with a documented case for why your scope justifies the top of it. Our guide to building a promotion case around strategic impact covers how to pair the market number with evidence of the work that earns it, which is what actually moves a comp committee.

For external leverage, remember that an outside offer isn't just a number—it's a process with its own timeline. Most competitive PM loops still open with a product sense conversation, and preparing for one properly, as our product sense interview framework lays out, takes real weeks, not days. Factor that lead time in before you assume an outside offer will be ready to use as leverage by a specific date.

It also helps to think of the negotiation itself as having its own arc—the ask, the pause, the counter, the silence, the close—rather than a single moment. Mapping that arc the way you'd map a customer journey for a product, complete with the points where anxiety spikes, helps you anticipate friction instead of being surprised by it mid-conversation.

Keeping Your Comp Research Ready for When You Need It

Research done for one raise conversation quietly evaporates by the next one—the screenshot gets lost, the peer conversation isn't written down, and eighteen months later you're rebuilding the same triangulation from scratch. The fix is treating comp research like any other reference material worth keeping, not a one-time scramble before a single meeting.

Prodinja's Library is built to hold exactly this kind of reference material—saved aggregator pulls, recruiter-quoted ranges, and notes from peer conversations—organized so they're retrievable rather than scattered across screenshots and old messages. When the actual conversation starts, the number you bring can be backed by the sources behind it instead of a half-remembered Reddit thread from months ago.

Key Takeaways

  • No single comp source is trustworthy alone: public aggregators skew toward brand names and active negotiators; recruiter ranges are built to close a hire; peer whispers are a small, biased sample. Triangulate all three.
  • Survivorship bias inflates what you see: people who negotiated well or landed prestigious offers are overrepresented in every public dataset, dragging the visible median above the true median.
  • Stage-mixing hides underneath one title: a "Senior PM" at Seed, Series C, and public-company stages can be three different jobs with three different cash-versus-equity mixes pooled into one misleading average.
  • Match scope to level before comparing anything—a title-to-title comparison across companies is comparing labels, not jobs; run your real scope through a leveling rubric first.
  • Adjust for stage and geography explicitly: equity-heavy early-stage offers and cash-heavy late-stage offers aren't directly comparable, and metro tier still meaningfully moves the number.
  • Bring a band, not a point: a defensible range anchored at the 50th percentile, with sources you can name, survives a follow-up question better than a single unsupported figure.
  • Keep the research alive between conversations—rebuilding your benchmark from scratch every raise cycle wastes the compounding value of what you've already learned.

Frequently Asked Questions

Is levels.fyi accurate for PM salaries?

It's directionally useful but skewed toward brand-name, tech-hub employers because submissions are voluntary and self-selected. Treat its ranges as a wide starting anchor to cross-check against recruiter-quoted ranges and peer conversations, not as a final number on its own.

How much should PM compensation vary by location?

Meaningfully, though the gap has narrowed with remote-first hiring. Tier-one hubs like the Bay Area and New York typically sit highest, secondary hubs like Austin or Denver land a step below, and fully distributed bands vary widely by company policy—so always ask which model a specific employer uses.

Should I share my current salary with a recruiter?

Generally, no—share a target range built from your triangulated research instead of your current number, since anchoring to what you already make (especially if you're underpaid) can cap the new offer below what the role and your level actually warrant.

How often should I refresh my comp benchmark?

At least once a year, and always before a raise conversation, a performance review cycle, or accepting a new offer—comp bands shift with the broader hiring market faster than most people update their mental model of "market rate."

What's a defensible total comp range for a Senior PM?

There's no single universal number—it depends entirely on level, company stage, and metro, which is exactly why triangulation matters more than memorizing one figure. Build your own band using the aggregator-plus-recruiter-plus-peer method above rather than anchoring to a headline number from a single source.

For the broader arc of leveling up your PM career—titles, promotion timing, and comp all included—see the complete guide to advanced PM career strategy.