Compa-Ratio: What It Is and How to Calculate It
Compa-ratio compares a salary to the midpoint of its range. The formula, worked examples, what a good ratio is, and what to do when the number is off.
Compa-Ratio
The metric that tells you whether somebody is paid fairly for their role, the formula, and the problem nobody mentions: you probably do not have a midpoint yet
Every article about compa-ratio gives you the formula, and the formula takes ten seconds. Salary divided by midpoint, times a hundred. Done. And then you go to apply it to your own team and stop dead, because you do not have a midpoint, and nothing you just read mentioned that this might be an issue.
That is the actual problem with compa-ratio at a small business. Not the arithmetic. The arithmetic is trivial. The problem is that the metric requires a salary structure to exist, and most businesses of five to fifty people have never built one, which means the number is uncomputable and nobody has told you.
So this page does the formula, quickly, and then spends most of its time on the part that matters: where to get a midpoint when you do not have one, how to read the number once you have it, and what to actually do when it comes out wrong. I build FirstHR, which is where compensation data lives. One caveat: pay decisions have legal consequences, and I am not a lawyer.
What Is Compa-Ratio?
Compa-ratio compares what you actually pay someone to the midpoint of the salary range for their role, expressed as a percentage.
The purpose is diagnostic. Compa-ratio does not tell you what to pay somebody. It tells you where they currently sit relative to what you decided the job is worth, and it prompts a question: is that where they should be, and can you say why?
Note the word decided in that sentence. The midpoint is not a fact about the world. It is a number you chose, and the quality of your compa-ratio is entirely dependent on the quality of that choice.
The Formula
Here it is, and this is the part every other article stops at.
An employee earning $54,000 against a midpoint of $60,000: 54,000 divided by 60,000 is 0.90, times 100 is a compa-ratio of 90. They are paid 90 percent of the midpoint.
You will sometimes see it expressed as a decimal instead: 0.90 rather than 90. Same number, different convention. A compa-ratio of 0.75 and a compa-ratio of 75 mean identical things, and which one you see depends entirely on who wrote the spreadsheet.
That is genuinely the whole calculation. Which is why the rest of this page is about the number you divide by.
One definitional note before we go further: the salary in the numerator is base salary, annualized. Not gross pay including overtime, not total compensation including benefits. If you are inconsistent about this across people, the comparison stops meaning anything, and the distinction between those figures is covered in gross pay versus net pay.
You Do Not Have a Midpoint
Here is the section that does not exist anywhere else on this topic, and it is the reason most people who read about compa-ratio never actually use it.
Compa-ratio requires a salary range to exist. Not a salary. A range: a minimum, a midpoint, and a maximum, defined for each role. And if you run a business with fifteen people and you have never sat down and built pay bands, then you have no midpoints, and compa-ratio is not a metric you can compute. It is a metric you can read about.
That accumulated residue has a name and a consequence. Undocumented pay decisions, made individually and never written down, are exactly how pay inequity arises at a small business, and it arises without anybody intending it.
This is not a reason to give up. It is a reason to reframe. If you go looking for compa-ratio and discover you cannot calculate it, you have just learned something more valuable than the metric would have told you: that you have no pay structure, and that every compensation decision you make is currently being made without a reference point.
Building the range is the work. The ratio is a byproduct of having done it. And the good news is that building one is genuinely an afternoon, not a project, which is the next section.
Where to Get One
You need to know, for each role, roughly what the market pays a fully competent person to do it. Here is where that number comes from, in descending order of how much I would trust it.
The Bureau of Labor Statistics occupational wage data is free, official, and the right place to start when you have nothing. It gives you median wages by occupation and by metropolitan area, which means you can get a defensible anchor for most roles without paying anybody. It is broad rather than precise, and broad is considerably better than nothing.
Occupation definitions and the tasks each role actually involves are on O*NET, which is useful when you are trying to work out which BLS occupation your role actually maps to, and that mapping is where most people go wrong.
Job postings are the underrated source, and they have become much more useful recently. A growing number of states now require employers to publish pay ranges in job advertisements, which means the market is telling you what it pays, in public, for free. Collect ten postings for the same role in your area, look at the ranges, and you have a rough picture in an afternoon.
A word on the roles themselves. If your job titles are invented rather than descriptive, you cannot map them to any external wage data, which means you cannot build a band. A clear job description is the prerequisite to all of this, and writing one properly is both better hiring and the thing that makes this measurable.
Once you have a midpoint, the range is usually built around it: a minimum somewhere around 80 percent of midpoint, a maximum around 120 percent. That is the conventional shape, it is arbitrary, and it is fine. What matters is that the range exists and that you can say where it came from.
How to Calculate It
Individually and for a group, worked all the way through.
The individual calculation is one line. The group calculation has two methods and the difference between them matters more than it looks.
The average method takes each person's compa-ratio and averages them. It treats everybody equally: the ratio of the person earning $40,000 counts exactly as much as the ratio of the person earning $140,000. This is what you usually want, because the question you are asking is are my people paid fairly, and every person is one person.
The aggregate method divides total salaries by total midpoints. It weights the result toward higher earners, because their salaries are larger numbers. This answers a different question, which is where does my payroll sit relative to market, and it is more of a finance question than a fairness question.
At your size, use the average method. With a small team and a wide salary spread, the aggregate method can be almost entirely determined by one or two senior people, which is exactly the distortion you do not want when you are trying to find out whether somebody junior is being underpaid.
What Is a Good Compa-Ratio?
The conventional answer is 90 to 110 for a fully competent employee. The more useful answer is: it depends on who the person is and how long they have been doing the job.
Read those bands as questions rather than verdicts. A compa-ratio of 82 is not a problem. It is a prompt. It becomes a problem only when you ask why and cannot answer, or when the answer turns out to be something you would rather not say out loud, like she did not negotiate as hard.
The distinction that actually matters is tenure and competence, not the number itself. A person three months into a role at a compa-ratio of 85 is exactly where they should be. The same person, three years in, fully competent, still at 85, is a resignation waiting to happen, and you will be surprised by it, and you should not be.
Note also that the metric only sees base salary. If a chunk of somebody's compensation arrives as commission or a bonus, the ratio is describing a fraction of what they actually earn. That category is supplemental pay, and it is invisible here.
The single most useful discipline here is not hitting a target ratio. It is being able to explain every number you see. If you can say, for each person, why they sit where they sit, and the reasons are job-related, you are in good shape regardless of what the distribution looks like.
Which means the reasons need to be recorded somewhere at the moment you make the decision, rather than reconstructed years later when somebody asks. That belongs in the personnel file, and a reconstructed rationale reads exactly like what it is.
What to Do When It Is Off
You ran the numbers and something looks wrong. Here is what each finding actually means and what to do about it.
Two rows there need emphasis. Above 120 with an average performer: freeze, do not cut. You overpaid at hire or the market moved, and the instinct is to correct it downward. Do not. Cutting somebody's pay is a different and much worse problem than overpaying them, it destroys trust immediately, and in some circumstances it creates legal exposure. Hold their pay flat and let the band catch up over time.
The employer's obligations here are set out by the EEOC on compensation discrimination, and they are stricter than most owners expect.
And the last row, which is the one to take most seriously. If the low compa-ratios cluster by demographic, stop. You are no longer looking at a compensation exercise. You are looking at a potential pay equity problem, and it has legal consequences that a spreadsheet will not solve. Note in particular that under the Equal Pay Act you may not remedy such a gap by reducing the higher-paid employee: the only lawful fix is to raise the person who was underpaid.
Compa-Ratio vs Range Penetration
The same salary, measured against two different reference points. They answer subtly different questions and it is worth knowing which one you are asking.
| Compa-ratio | Range penetration | |
|---|---|---|
| Measured against | The midpoint of the range | The full span from minimum to maximum |
| Formula | (salary / midpoint) x 100 | ((salary - minimum) / (maximum - minimum)) x 100 |
| At the midpoint | 100 | 50 percent |
| At the minimum | Depends on how wide the band is | 0 percent, always |
| At the maximum | Depends on how wide the band is | 100 percent, always |
| Most useful when | Comparing across roles with similar band widths | Your ranges vary a lot in width |
The key relationship, and the thing worth memorizing: an employee at the exact midpoint has a compa-ratio of 100 and a range penetration of 50 percent. Both describe the same person sitting in the same place. Compa-ratio measures distance from the middle. Range penetration measures how far up the band they have climbed.
If you are running this exercise because turnover has become a problem rather than because you are curious, the metric is downstream of the diagnosis. Whether pay is actually why people are leaving is a question worth answering before you spend money on it, and incentive programs are one of the alternatives.
Range penetration is more sensitive to band width. If one of your ranges is narrow and another is wide, the same compa-ratio of 110 means quite different things in each, whereas range penetration normalizes for that automatically. For most small businesses with a handful of roles, compa-ratio is the simpler and sufficient tool.
What It Does Not Tell You
Worth being honest about, because the metric is frequently presented as more powerful than it is.
| What compa-ratio does not know | Why it matters |
|---|---|
| Whether your midpoint is any good | The entire metric is relative to a number you chose. A bad midpoint produces confident, precise, wrong answers |
| Anything about performance | It compares salary to a band. It has no opinion on whether the person is any good, and it should not be read as if it does |
| Total compensation | Bonuses, equity, and benefits are all invisible to it. Two people at the same compa-ratio can be paid very differently in reality |
| Whether the role is correctly defined | If somebody has outgrown their job description, the band is measuring the wrong job and the ratio is measuring nothing |
| Why the number is what it is | It surfaces a gap. It does not explain it, and the explanation is the entire value of the exercise |
| Whether a pattern is discriminatory | It can show you a cluster. Interpreting that cluster is a legal question, not a mathematical one |
The third row is worth a note too. Compa-ratio sees base salary and nothing else, which means the full cost of employing somebody, including the employer taxes and benefits covered in statutory benefits, is entirely outside its view.
The first row is the one to hold onto. Compa-ratio inherits every flaw in your midpoint. If you built the band from your own existing salaries, every ratio will land near 100 and you will conclude that everything is fine, when what you have actually done is measure your salaries against themselves.
And the second row deserves saying plainly, because the metric gets misused this way constantly: compa-ratio is not a performance rating. A person with a compa-ratio of 92 is not worse than a person at 108. They are paid differently, and the reason might be tenure, might be scarcity, might be that one of them negotiated better on a Tuesday in 2023. The number does not know and does not care.
Is This Worth It at Your Size?
An honest answer, since most content on this subject is written by people who sell compensation software.
| Your situation | Is compa-ratio worth it? | What to do instead, or as well |
|---|---|---|
| Under 10 employees | Not really | You can hold every salary in your head. But write down a range per role anyway, because the discipline is what matters |
| 10 to 25 employees | Starting to be | This is where you stop being able to see the whole picture at once. Build the bands. The ratio follows |
| 25 to 50 employees | Yes | You now have enough people that patterns exist and you cannot see them without a metric |
| You have no salary ranges | You cannot calculate it at all | Build the ranges. That is the work, and the metric is a byproduct |
| You are about to raise or sell | Yes | Diligence will ask about your comp structure. Not having one is a finding |
| You suspect a pay equity problem | Yes, urgently | But treat it as a legal question rather than a comp question, and get advice |
For the market data itself, the BLS National Compensation Survey is the other free federal source worth knowing about, and it covers benefits as well as wages.
The recurring answer in that table is the same one: the salary ranges are the valuable thing, and compa-ratio is what you get for free once they exist. If reading this page produces nothing except an afternoon spent building bands for your roles, it has done its job, and you can ignore the metric entirely.
Wherever this data lives, it needs to live somewhere. Salaries, effective dates, and job classifications scattered across offer letters and somebody's memory is precisely why this exercise feels impossible, and it is what an HRIS exists to fix.
The underlying question is worth asking at any size, and it is not a mathematical one: is each person paid roughly what their role is worth, and can I explain every difference? At five people you can answer that in your head. At twenty-five you cannot, and that is precisely where a number starts earning its keep. Where this fits alongside everything else worth measuring is covered in HR metrics.
Common Mistakes
Six recurring errors, and the first one invalidates everything downstream of it. If you have never examined your compensation history at all, an HR audit is the structured way to find out what is in there while it is still cheap to fix.
Frequently Asked Questions
What is compa-ratio?
Compa-ratio, short for comparative ratio, is a compensation metric that compares an employee's actual salary to the midpoint of the salary range for their role. It is calculated by dividing the actual salary by the midpoint and multiplying by 100, producing a percentage. A compa-ratio of 100 means the employee is paid exactly at the midpoint. Below 100 means they are paid below it, and above 100 means above it. It is used to check whether individuals are paid fairly relative to the range you set for their job, and to spot patterns across a team.
What is the compa-ratio meaning in simple terms?
It is a way of asking whether somebody is paid roughly the right amount for the job they do. You take what you actually pay them, divide it by the middle of the range you think that job is worth, and turn it into a percentage. If they earn $54,000 and the middle of the range is $60,000, their compa-ratio is 90, meaning they are paid 90 percent of the midpoint. The number itself is not a verdict. It is a prompt to ask why.
How do you calculate compa-ratio?
Divide the employee's actual salary by the midpoint of their salary range, then multiply by 100. An employee earning $54,000 against a midpoint of $60,000 has a compa-ratio of 90. For a group, you have two options: average the individual compa-ratios, which treats every person equally, or divide the sum of all salaries by the sum of all midpoints, which weights the result toward higher earners. For a small business, the average method is usually the more useful of the two.
What is a good compa-ratio?
Conventionally, 90 to 110 is considered healthy for a fully competent employee, meaning they are paid at or near the midpoint of their range. Below 90 is normal for new hires still developing in the role. Above 110 typically indicates a strong performer, scarce skills, or long tenure. Below 80 or above 120 are worth investigating. But these bands are conventions rather than rules, and a number outside them is a question rather than a problem. The discipline that matters is being able to explain every number you see.
What does a compa-ratio of 90 mean?
The employee is paid 90 percent of the midpoint of their salary range, so 10 percent below the middle of the band. That is entirely normal for somebody who is relatively new to the role or still developing full competence in it, and it is a reasonable place for them to sit while they grow into the job. What would be a problem is an employee who has been fully competent for two years and is still at 90, because that suggests their pay has not kept pace with their capability.
What does a compa-ratio of 0.75 mean?
The same thing as 75, just expressed as a decimal rather than a percentage. The employee is paid 75 percent of the midpoint, meaning 25 percent below the middle of the range for their role. That is materially below the band and it is worth understanding why. If the person is genuinely new and still learning, it may be defensible for a period. If they are performing the job fully, it is a real gap and it is the kind of gap that ends in a resignation you did not see coming.
Where do I get the midpoint?
This is the actual hard part, and most articles about compa-ratio skip it entirely. The midpoint is the middle of the salary range you have set for a role, which means you need a salary range first. If you do not have one, you cannot calculate compa-ratio at all. Free sources for building one include Bureau of Labor Statistics occupational wage data and, increasingly, job postings for the same role, since a growing number of states require employers to publish pay ranges. What you should not do is anchor the midpoint on what you already pay, because that just enshrines whatever you are currently doing.
What is the difference between compa-ratio and range penetration?
They measure the same salary against different reference points. Compa-ratio measures a salary against the midpoint of the range. Range penetration measures where a salary falls between the minimum and the maximum, expressed as a percentage of the full range. An employee at the exact midpoint has a compa-ratio of 100 and a range penetration of 50 percent, because they are halfway between the bottom and the top. Range penetration is more sensitive to how wide the band is, which makes it useful when your ranges vary a lot in width.
How do I calculate group compa-ratio?
Two methods. The average method takes the individual compa-ratio of each person and averages them, which treats every employee equally regardless of what they earn. The aggregate method divides the sum of all salaries by the sum of all midpoints, which weights the result toward higher earners because their salaries are larger. Both are valid and they answer slightly different questions. For a small team, the average method is usually more informative, because the aggregate method can be dominated by one or two senior salaries.
Is compensation ratio the same as compa-ratio?
In an HR context, yes: compensation ratio, comparison ratio, and compa-ratio are used interchangeably to mean the same metric. Be aware, though, that outside HR the phrase compensation ratio has a completely different meaning in finance, where it refers to compensation expense divided by revenue and is used to assess how much of a firm's income goes to paying its people. If you are reading something about compensation ratios in the 30 to 40 percent range, you are reading about the finance concept, not this one.
Is compa-ratio worth tracking at a small business?
Only once you have salary ranges, and the ranges are the valuable part rather than the metric. Compa-ratio is a diagnostic that runs on top of a pay structure, so if you do not have a structure, building one is the work and the ratio is a byproduct. That said, the underlying question is worth asking at any size: is each person paid roughly what their role is worth, and can I explain every difference? At five employees you can answer that in your head. At twenty-five you cannot, which is roughly where the metric starts earning its keep.
Can compa-ratio reveal pay discrimination?
It can surface a pattern, and a pattern is what matters. A single low compa-ratio is a question about one person. Low compa-ratios that cluster by sex, race, or another protected characteristic are not a compensation question at all; they are a pay equity question with legal consequences. If you run the numbers and the low ones group by demographic, stop treating it as a comp exercise and treat it as what it is. Note also that under the Equal Pay Act you may not fix such a gap by cutting anybody's pay: the only lawful remedy is to raise the underpaid employee.