What is salary benchmarking?
And how to run one properly.
Salary benchmarking, also known as compensation benchmarking, is the process of comparing the pay for roles inside an organisation against external market data for equivalent roles. It establishes whether current pay sits above, at or below the market, and produces the ranges an employer uses to set offers and plan review cycles.
Why organisations benchmark pay
Four pressures put benchmarking on the agenda, and most organisations arrive through one of them rather than through all four at once.
- Attracting people. An offer priced below the market loses the candidate late, usually after the interviewing is done and the cost is sunk.
- Keeping them. Pay drifts quietly. Someone hired three years ago at the market rate is often below it now, and the person most likely to notice is a recruiter calling them.
- Planning the budget. A defensible range turns a salary bill from a set of individual negotiations into a number that can be forecast.
- Pay transparency and equity. Disclosure rules increasingly require published ranges, and a published range needs something behind it. Benchmarking is also how unexplained gaps between comparable people surface before anyone else finds them.
How the process works
A benchmarking exercise runs in four steps. The second is where most of them go wrong.
- Collect the data. Published salary surveys, industry reports, aggregator data and, where it can be had, live evidence from people currently in the role. Each source has a different bias, which is why more than one is used.
- Match the roles. Compare jobs by the work they actually contain rather than by title. Two companies can both employ a head of operations and mean entirely different jobs; matching on the title alone produces a confident answer to the wrong question.
- Adjust for the comparators that move pay. Industry, company size, funding stage, scope of the role and the market you are hiring from. A number unadjusted for these is an average that takes in companies outside your talent market.
- Set the ranges. A minimum, a midpoint and a maximum for each level, with a stated position against the market. Paying at the median is a decision; so is paying at the seventy-fifth percentile for scarce roles and the median elsewhere.
The step organisations skip is the second one, because it is slow and unglamorous. We have done it badly ourselves on a rushed mandate and produced a range that was accurate for a job nobody was hiring for.
Where the data comes from, and what each source misses
Every source gives a partial picture. Understanding what each one leaves out is most of the skill.
| Source | Strength | What it misses |
|---|---|---|
| Published salary surveys | Large samples, consistent method year to year | Lags the market by six to twelve months |
| Aggregator and self-reported data | Current, broad, cheap to reach | Self-selection, and base pay reported separately from the variable elements |
| Recruiter and search firm data | Reflects what offers are actually closing at | Skewed towards roles that firm recruits for |
| Practitioners in the seat | Explains the number, including the parts a survey flattens | Small samples, gathered one conversation at a time |
Marquee reads survey data against evidence from practitioners currently doing the work, which is what the Nextyn network exists to provide. A survey tells you the range. A person in the role tells you why the top of it is being paid, and whether the figure includes a bonus that rarely pays out.
Benchmark total compensation, rather than base salary
Base salary is the easiest element to compare and the least informative on senior roles, where it is often the smaller half of the package.
- Annual bonus, and crucially the percentage that actually pays out against target.
- Equity, its vesting schedule, and whether it carries real liquidity.
- Pension contribution, which varies enough between employers to move the total several percentage points.
- Allowances, notice period and severance terms, all of which have a cash value to the person weighing the offer.
A candidate on a lower base with a reliable bonus and a funded pension can be better paid than one with a higher headline number. Benchmarking base alone produces offers that look competitive on paper and lose people in the detail.
Reading the output: compa-ratio and range penetration
Two measures turn a set of ranges into something a manager can act on.
Compa-ratio
A person’s salary divided by the midpoint of their range, expressed as a percentage. A compa-ratio of 100 sits exactly at the midpoint. Between 80 and 120 is the usual working band: below 80 suggests someone underpaid or newly promoted, above 120 suggests someone at the ceiling of their level whose next raise is a promotion rather than an increment.
Range penetration
Where a salary sits between the minimum and maximum, as a percentage of the whole range. It answers the question a compa-ratio leaves open, which is how much room is left before the level runs out.
Deciding where to sit against the market
Benchmarking produces a distribution. Choosing a position within it is a separate decision, and it is the one that actually sets the salary bill.
- At the median. Half the market pays more. Workable where the employer offers something else people want, and expensive in attrition where it offers only the salary.
- At the upper quartile. Pays a premium to shorten time to hire and hold onto people. Justifiable for roles where a vacancy costs more per week than the uplift costs per year.
- Below the median, deliberately. A real strategy for organisations with strong non-cash draws, provided it is chosen openly rather than arrived at by drift.
The common error is applying one position everywhere. A single percentile across the whole organisation overpays for roles with a deep candidate pool and underpays for the two or three that are genuinely scarce, which are usually the ones causing the problem that prompted the exercise.
A differentiated position is more defensible than it sounds, provided the logic is written down. Scarcity, business criticality and how quickly the role could be replaced are all reasonable grounds to pay one function above another. What creates difficulty is paying differently and being unable to say why.
The limits of a benchmark
A benchmark describes what the market pays. It stays silent on the three things that most often decide an offer, and treating it as a complete answer is how organisations end up confidently underpaying.
- What this particular person is worth to you, which depends on scarcity of the skill and what stalls while the seat is empty.
- What they will accept, which is shaped by their notice period, unvested equity and how content they are where they sit.
- What your competitor will offer if it comes to that, which sits beyond the reach of any survey.
The range is the starting position for a decision, rather than the decision. Where the two conflict, the market evidence is usually right about the middle of the distribution and wrong about the edges, and senior roles live at the edges.
How often to benchmark
A full exercise annually, timed to land before the pay review rather than after it, is the common cadence. Scarce or fast-moving functions justify a lighter check every six months.
Two events force it regardless of the calendar: opening a role you are hiring for the first time, where the organisation lacks an internal reference point, and losing two people from the same level in quick succession, which is usually pay telling you something before anyone says it out loud.
Frequently asked questions
To establish whether internal pay is competitive against the external market, and to produce defensible ranges from that comparison. Everything downstream follows from it: offers that close, review cycles that can be budgeted, and published ranges with something behind them.
They are the same exercise under two names. "Salary benchmarking" is the more common phrasing in the UK and Europe; "compensation benchmarking" is more common in the United States and tends to signal that bonus and equity are in scope alongside base pay.
Around 100, meaning the salary sits at the midpoint of its range. Most organisations work within a band of 80 to 120. Below 80 usually means someone underpaid or recently promoted; above 120 means someone near the ceiling of their level, where the next move is a promotion rather than a rise.
It ranges from a few hundred for a single-role check to five figures for a full-organisation study. Cost tracks the number of roles, the number of markets, and whether the work relies on published survey data alone or adds primary evidence from people currently in those roles.
The right answer depends on how senior and how scarce the roles are. Aggregator platforms cover high-volume roles well and cheaply. Senior and specialist roles have samples too small for any platform to be reliable, and those are where primary evidence earns its cost.
Annually for most organisations, timed to land before the pay review. Every six months is justified in fast-moving or scarce functions. Opening an unfamiliar role, or losing two people from one level in quick succession, forces the question regardless of the cycle.
It should, and on senior roles it has to. Base salary is often the smaller half of a senior package. Benchmarking base alone produces offers that read as competitive and lose candidates once the bonus target, vesting schedule and pension contribution are compared.
