There Is No "Market Rate." There's a 2x Range and Where You Land Is Politics.

"Market rate" is one number doing the work of a hundred decisions. For US software developers, the same job title in the same country pays $79,850 at the 10th percentile and $211,450 at the 90th, a 2.65x spread per the Bureau of Labor Statistics. That gap is not skill noise. It is which company, which band, and how hard you pushed. So "am I being paid market rate?" is the wrong question.

Here's the thing nobody tells you when you Google your salary: the average you find is a blend of people who got crushed and people who got paid, all wearing the same job title. Treating that blend as your "fair number" anchors you to the middle of a range you could be climbing. The number you should chase isn't the average. It's the ceiling of the specific band you're standing in.

Why doesn't a single "market rate" actually exist?

Because no single market sets it. Pay is the sum of internal calls each company makes: which salary survey they bought, what percentile they decided to anchor to, how wide they drew the band, what level they slotted you at on hire, and whether they flex for a counter-offer. None of that is a market clearing one price. It's a fragmented stack of policies.

The BLS proves the spread is structural. Software developers run $79,850 (P10) to $211,450 (P90). Financial and investment analysts run $62,410 to $180,550, a 2.9x spread. Same occupation. Same country. Same Census of titles. When the bottom and top of one job differ by nearly 3x, "the market rate" describes a herd, not a price. You don't get paid the average of a herd. You get paid wherever your specific company puts you in their specific band.

What sets the ceiling before you even interview?

The company's pay philosophy, decided long before your resume showed up. This is the same reason most of your negotiation outcome is locked in before you say a number: the band and the percentile were chosen without you in the room. Most companies are middle-of-the-pack on purpose. Ravio's analysis found 77% of tech companies sit at the 50th percentile of the market, and only 9.6% target the 75th or above. Netflix is the famous outlier: it explicitly targets the 90th percentile, "personal top of market," with a 2023 median employee salary of $200,761. Deep tech is its own world: 50% of deep tech firms pay at the 75th percentile or above.

So before you say a word in the interview, the company has already picked which list it shops from. Someone Googling "market rate for software engineer" gets an average across all of it. That average isn't their ceiling at a deep tech firm, and it isn't their floor at the e-commerce company where a chunk pay below median. Same search. Wildly different reality, depending on whose door you walked through.

Weak version: Two senior software engineers, same city, same LinkedIn title. One's at a Series B fintech anchored to the 50th percentile, around $200K total comp. The other's at Netflix, where a Level 5 engineer's base salary alone ran $423,700. On paper they have identical jobs.

Strong version: Neither one is "underpaid." They're in different bands at different companies that chose different percentiles. The fintech engineer isn't bad at negotiating. They picked, knowingly or not, a market that tops out lower. The 2x gap is a company-selection decision wearing a salary-negotiation costume.

What's the real unit of pay, if not the title?

The band. Every structured company groups roles into pay bands, each with a midpoint, and lets actual pay range around it. The standard acceptable spread is a compa-ratio of 80% to 120% of midpoint. Do the math: two people with the same title in the same grade at the same company can sit at 80% and 120% of the same midpoint. That's a legal, normal, HR-approved 50% gap. Neither is "wrong" on paper. It's also why, when you push for more, the answer "we can't go higher" is usually about that band, not the budget.

Titles make it worse, because levels don't transfer. A mid-level engineer (L4) at Google earns a $305K median total comp, while a "Senior Engineer" at a non-tech company might pull $120K. The word "Senior" is doing no real work across companies. Google's own ladder runs L3 $214K, L4 $305K, L5 (Senior) $425K, L6 (Staff) $620K. The level, not the title, is the unit. And the level is an internal artifact that means nothing the moment you leave.

This is also why hopping for a title is a trap and hopping for a band is a move. A bigger word on your LinkedIn that lands you in a lower band is a downgrade you'll feel for years. The story that matters is the one where each move walks you up a band, not up a font size.

Where in the band do you actually land, and what decides it?

Politics and timing, mostly. Two new hires, same role, same grade: one starts at 80% of midpoint because they were a nervous new grad who took the first number, the other at 120% because they walked in with a competing offer. The "internal equity" rule only stops the most extreme outliers. It does not collapse the band. The band is wide on purpose, and you get slotted into it based on bargaining power, not merit.

The evidence on what that bargaining power is worth:

What you did	What happened
Negotiated	85% got at least part of the ask
Negotiated	18.83% average gain vs. those who didn't
Took the first offer	46% of candidates do this
Employer's expectation	73% expect you to negotiate anyway

And the default landing spot is low, not middle. Glassdoor matched 147,568 job listings to real salaries and found actual pay skews toward the lower half of the posted range, with 22% of postings paying below the range entirely. The company posted $140K to $200K. Most people landed near $155K. The $200K was sitting right there in the posting. Nobody asked for it.

What three questions replace "what is my market rate?"

Three questions that have real answers, where "market rate" has none.

1. Which band am I in? Find the midpoint and the range for your level at your company, or your target company. Public salary-range laws in several US states and crowdsourced data make this more findable than it used to be, though those transparency rules ended up helping employers at least as much as candidates.
2. Which percentile does this company target? Median (most of them), 75th, or 90th. This sets your ceiling. A 50th-percentile company will not pay you a 90th-percentile number no matter how well you interview.
3. What's the top of this band? Not the "fair" midpoint. The top. Senior engineers in the US run from a $176K (P25) floor to a $480K (P90) ceiling on Levels.fyi. Aim at the ceiling of your real band, then negotiate toward it.

Notice what these have in common. They're answerable. You can look up a band, a company's pay philosophy, and a ceiling. You cannot look up "the market rate," because it's an average pretending to be a target.

How do you actually use this in your next offer?

Three moves, in order.

First, identify the company's pay tier before you fall in love with the role. Is it a median-targeting company (the 77% default) or a 75th-percentile-plus outlier? That tells you the shape of the whole conversation. A median company's "great offer" is a different animal from a deep tech firm's.

Second, find where you sit in their band. Use Levels.fyi and Glassdoor filtered to that company and level, not the generic role average. The generic average blends the e-commerce floor with the Netflix ceiling and tells you nothing about this specific door.

Third, negotiate to the top of the band, not to the "fair number." The fair number is the midpoint, and the data says you'll drift below it if you do nothing. Asking is normal. 73% of employers expect it. The ceiling is real budget, already approved, sitting in the band. You're not being greedy. You're declining to leave approved money on the table.

What does this cost you?

Three things, plainly.

You'll do more homework. Finding the band, the tier, and the ceiling takes a few hours of research per offer instead of one Glassdoor glance. That's the price of not anchoring to a blended average.

You'll sometimes pick the lower-band company on purpose. If you won't interview at the 90th-percentile firms, the 50th percentile is your honest ceiling, and that's a legitimate choice. The trade-off isn't that median is a lie. It's that calling it "the market rate" hides the choice you're making about which market to stand in. Maybe the median company has the team, the commute, the calm you actually want. Name what you're paying for and pay it on purpose.

And the negotiation gains have an asterisk. People who negotiate are often more prepared candidates to begin with, so the 18.83% headline likely overstates the pure causal effect. A more conservative read is that even a 5-10% bump is normal for asking, and over a career that compounds into real money. The point holds: asking moves you up the band you're already in.

What about the obvious objections?

"The 2x spread is just different experience levels." Partly fair. The BLS P10-to-P90 doesn't control for seniority. But Levels.fyi senior-only data still runs $176K (P25) to $480K (P90), a 2.7x spread among people sharing one seniority label. Control for level and the gap barely moves.

"My recruiter said this is market." Your recruiter is quoting their company's band at the percentile their company targets. That's their market, not the market. It's a true statement about one band and a misleading one about your options.

"Leveling is meritocratic, I land where I belong." A senior bar at one company is a different bar elsewhere. The level reflects the company's internal demand and who happened to interview you, not your transferable value. Landing high in a band is part performance, part timing, part luck of the panel.

What to do now

Pick your next real offer or your current role. Look up three numbers, in this order: the company's pay-philosophy tier (median or above), your level's band and midpoint on Levels.fyi or Glassdoor filtered to that company, and the top of that band. Then set your ask at the ceiling and have the conversation. Most people accept the first number and land in the lower half of a range that had more room in it. You now know better. The money is in the band, already approved. Go ask for it.

Want to know which band you're actually in and what the ceiling looks like for your role and city before your next offer? Message Praxy on WhatsApp. We'll pull the real range together and build your ask.