Same Title, Different Pay: Inside the Company Pay Gap

At a software company in Denver last spring, the compensation director pulled the pay file for two engineers with the same title, the same level, and the same start year. One earned $128,000. The other earned $176,000. The first had been hired in a soft market and had collected routine 3 percent raises since. The second had been hired during the 2021 hiring surge, when the company paid whatever it took, and she had kept the ground she gained. On paper, the gap was 37 percent, for work the company itself rated identically.

The director, who has run pay audits at three employers and asked not to be identified, said the spreadsheet told a familiar story. "It is almost never one villain," she said. "It is a series of small decisions made years apart — a hire at market, a raise of 3 percent, a counteroffer, a promotion a year late. Each one is defensible on its own. The total is not."

The pattern is not confined to one Denver company. Compensation data from Payscale and Glassdoor show that the same job title at a single employer routinely spans pay gaps of 20 to 40 percent, and pay equity audits — the internal studies employers run to find unequal pay — keep surfacing the same result. What follows is how pay ranges work inside companies, why identical titles drift apart in pay, what the audits actually find, and who the data say ends up ahead. The question matters now because a growing number of states, including California and Colorado, are forcing employers to report the numbers.

How a title becomes a range

Most large employers do not pay people; they pay positions. Compensation teams price each job against market surveys from firms like Mercer and Willis Towers Watson, then build a band around the market midpoint — commonly 50 percent wide, with a minimum and a maximum. A product manager role with a midpoint of $110,000 typically spans about $88,000 to $132,000. The range is the employer's way of saying the title means something consistent.

Where someone lands inside the range is a different question, and it is where the dispersion begins. Compensation professionals measure position with the compa-ratio — pay divided by the midpoint. A compa-ratio of 1.0 means the employee sits exactly at market. Ratios of 0.8 to 1.2 are considered normal, which is another way of saying that two people doing the same job can sit 40 percent apart and still count as in range by every formal measure.

That design has a logic, and the logic is temporal. New hires are priced at the market of the day, which moves. Tenured employees are priced by annual raises, which barely move. When the market jumps — as it did in 2021 and 2022 — the band moves up, new hires enter near the top, and long-serving employees climb toward it at 3 or 4 percent a year. It can take a decade for raises to catch a market that moved in a year. The result is what compensation consultants call pay compression, and it shows up in audit after audit.

The discipline of maintaining all this varies. Well-run companies review bands against market surveys every year and adjust the people inside them. Many others update the bands and leave the people alone, which quietly moves everyone who stays below the midpoint. Title inflation adds a second layer: a senior analyst at one company is a manager at another, and the same label covers different levels of work, which is why ranges attached to titles are wider than the ranges attached to levels.

Two product managers, same title, same level, eight years apart

	Hired in 2016	Hired in 2022
Starting pay	$92,000	$118,000
Average annual raise	3.5 percent	3.5 percent
Pay in 2026	$121,000	$148,000
Compa-ratio	0.87	1.06

Where the gap comes from

Auditors who dig into the numbers find the same short list of causes, in roughly the same order. Hiring at market leads it: whoever joins when demand is hot enters the band high, and whoever joins in a slack year enters low, and the difference is never reconciled. Negotiation at entry follows. Research led by the economist Linda Babcock found that people who negotiate a first job offer add thousands of dollars to it, and studies of job changers suggest the advantage persists for years. Manager discretion comes third — most companies hand raises to managers with only loose guidance, and managers reward the people they notice.

The less visible causes matter as much. Retention counteroffers push one employee ahead of the band while everyone else stays put. Promotions often carry raises of 10 percent or so, so the timing of a promotion — a year early or a year late — becomes a permanent pay event. And because most raises are percentages, every early advantage compounds. A 5 percent edge at hire, carried through a decade of 3 percent raises, stays a 5 percent edge forever. Nothing about the work changes. Only the arithmetic does.

A recruiter in Austin who has placed software engineers for fifteen years described watching the same candidate priced differently by the calendar. "I have sent a profile in January and the same profile in June and the company moved the number by $15,000," he said. "Nothing about the person changed. The market did."

What the audits find

Pay equity audits are the employer's own ledger of all this. In a typical audit, a company hands its pay file — title, level, salary, tenure, performance rating, gender, race — to an outside firm, which runs a regression asking a blunt question: do people who differ only in gender or race earn different amounts for the same work? The answer, at most companies, is yes, though the size of the gap depends heavily on how the question is asked. Unadjusted — comparing everyone in a title regardless of tenure or performance — the gaps are largest, often 10 to 20 percent. Adjusted for experience and ratings, they shrink but rarely disappear, typically settling in the low single digits or just above.

The method matters, and companies choose it. Compare everyone who shares a title and the gaps look large; compare only within the same level and tenure cell and they shrink toward zero. Regulators push for the first approach, because it captures the structural part of the problem — who gets hired into the band, who gets promoted. Employers prefer the second. Both are true, and the choice decides the number.

A growing number of employers publish the results. Salesforce, which began auditing its pay in 2015, said it spent millions of dollars adjusting salaries and later reported that women and men at the company were paid within a fraction of a percent of each other for comparable work. Adobe said in 2018 that it had reached pay parity and has repeated the claim in most years since. Citigroup, which has published a global median gender pay gap since 2021, reported that women globally earn about three-quarters of men's median pay; the company says the gap reflects where women sit in the organization rather than unequal pay for equal work.

"The audit almost always finds the same thing," said a compensation consultant in Chicago who has run pay studies for more than twenty employers. "The people who joined in hot markets sit above the midpoint. The people who stayed sit below it. And the people who stayed are disproportionately women."

Who benefits, who pays

The most consistent finding in these audits is not about gender or race at all. It is about time. Workers who changed jobs at the right moment — in 2021, 2022, and parts of 2023 — carried market-rate pay into the band and kept it. Workers who stayed carried 2019 pay with 3 percent raises attached. The two groups occupy the same title, and Bureau of Labor Statistics wage data show why: job switchers' pay growth ran several points a year above stayers' through the early 2020s, and the gap has closed only partly since.

The same audits show who pays. Women and Black and Hispanic workers are overrepresented among long-tenured staff in many companies, which means the tenure penalty lands on them twice: once through the market timing, again through the demographic pattern. Economists who study within-firm pay have documented the effect of comparison itself. In a 2018 study of one company's records, researchers including David Card and Alexandre Mas found that workers' job satisfaction fell and quits rose when they learned that peers doing the same job earned more — even when the gap had a defensible explanation.

There is a case for dispersion, and employers make it. Scarcity deserves a price: a company that loses three engineers in a quarter will pay for the fourth. Performance deserves recognition: a top-rated employee at 1.05 compa-ratio is not the same problem as an average one at 0.85. The trouble is that the market does most of the deciding. Few of the same-title gaps that audits surface trace to performance ratings; most trace to the calendar.

The disclosure wave

The pressure to close these gaps has moved from internal audits to state law. California began requiring employers with 100 or more workers to file annual pay data by race, ethnicity, and sex in 2021. Colorado's Equal Pay for Equal Work Act, in effect the same year, went further, requiring pay ranges on every job posting and promotion notice. New York and Washington followed with posting laws of their own. The result is a growing public record of what companies pay, and researchers are beginning to read it.

Early findings from California's first filings showed the same structure that internal audits reveal: pay gaps concentrated at the top of organizations, with women and people of color underrepresented in the highest-paid roles and overrepresented in the lowest. Studies of posting laws, summarized by the Economic Policy Institute, have found that published ranges narrow the spread of offers and shift negotiating power to candidates, who no longer have to guess. The laws have also changed behavior inside companies, compensation consultants say — a manager who knows a range will be posted tends to price new hires more carefully.

The California data, collected from more than six thousand employers, gave researchers their first look at the inside of the pay file at scale. Early analysis of the filings found the same shape the audits show: gaps that track where people sit more than what they are paid for identical work — the sorting problem again, now measured on a statewide scale.

The limits are just as visible. Posting laws govern what is advertised, not what is paid; internal gaps can persist even when every range is public. State reporting captures snapshots, not explanations, and enforcement is uneven. Companies have also found workarounds — broad ranges that span 50 percent or more, titles tailored to dodge comparisons, and performance ratings that retroactively justify the numbers.

What workers can do

For the employee sitting at the bottom of a band, the first step is knowing where the band is. Salary ranges are less secret than they once were: Payscale and Glassdoor publish market data by title and city, and state posting laws mean many ranges are public. Our salary calculators answer the same question from the other end: what is the number, in this market, for this work. From there, the conversation changes from a request to a comparison.

That comparison is the core of the modern ask. A worker who can say "the midpoint for this title is $118,000 and I am at $104,000 — a compa-ratio of 0.88" has stopped asking for a favor and started documenting a discrepancy. Consultants say the most effective conversations name a number, cite a source, and propose a date. Timing matters too: raises are decided in budget season, usually in the fall, and a request made after the budget is set is a request deferred.

The script does not need to be long. The Chicago consultant teaches clients a three-sentence version: "I'd like to talk about where my pay sits relative to the range for this role. I understand the midpoint is $118,000 and I'm at $104,000. Can we look at what it would take to bring me to market?" It works, she says, because it is a comparison, not a plea — and because it forces the manager to answer a question the company has already asked itself in the audit.

The harder question is whether to stay and fight or leave and reset. The data are blunt: switchers have out-earned stayers by several points a year since 2021 — Federal Reserve Bank of Atlanta wage-tracker data — and the premium, though smaller, has not disappeared. An offer in hand is still the strongest argument an employee can make, and it is the one argument an internal-equity review cannot touch. The consultants who run audits offer a caution: a company that has already measured its gaps knows what it would cost to close yours, and it has already decided.

What employers can do

Employers that take the audits seriously tend to do the same four things. They fix the outliers — the top and bottom of the distribution — with one-time adjustments. They tighten hiring ranges so new hires cannot enter above midpoint without sign-off. They give long-serving employees market adjustments on top of merit raises. And they repeat the audit every year or two, because the gaps regenerate. The Society for Human Resource Management, which advises companies on these reviews, says the pattern is predictable: fix, drift, audit, fix.

The economics push in the same direction. A pay-equity fix is a one-time expense, usually a small fraction of payroll; a quiet exodus of tenured staff is a recruiting cost that compounds. Research on quitting shows that pay comparisons, not just pay levels, drive departures — the Card and Mas findings again — which means a visible gap is itself a retention risk. Companies that disclose ranges and publish audit results are, in effect, choosing which risk to buy.

Not every company plays this game. Some employers still run audits quietly and adjust nothing, betting that secrecy will hold. The consultants say the bet is getting worse. Each new state law publishes another company's ranges, each job board exposes another title's spread, and the worker who once had no way to know now has three.

The companies that refuse are increasingly the exception, and the workers they employ have the clearest remedy of all. A job market that prices titles publicly — through posted ranges, salary sites, and state data — has made the exit the default answer to a stalled negotiation. The consultants say the pattern of the last five years is that the worker who documents the gap either gets the adjustment or gets the offer.

What to watch

The direction of travel is clear. The European Union's pay transparency directive, which took effect in 2026, requires large employers to report gender pay gaps and publish starting salaries, and American states keep adding their own versions. The federal government has twice moved toward collecting pay data from large employers, and the reporting requirements, whatever their legal fate, have made the question permanent. Pay equity audits are no longer a correction a company makes once; they are a line item in the annual budget.

For the worker, the practical question is narrower: where do you sit, and can you document it? The tools exist — market data from Payscale and Glassdoor, posted ranges from the states that require them, and calculators like the ones on this site that turn an annual number into a comparable one. A title is no longer a guarantee of anything except a label. The pay behind it is negotiable, measurable, and increasingly public.

The Denver engineers who opened this story are, by now, part of a pattern their company has already priced. The audits will keep finding the gaps; the question is which companies close them before the workers do the closing themselves — by leaving. Every year of new data makes that exit easier.