Salary survey cost: what compensation survey data costs, and when benchmarking software is cheaper

Almost nobody in compensation data publishes a price. The clearest published figure in the whole category is Mercer MarketPricer at 300 USD per job. For full survey subscriptions you are into quote-only territory, where third-party 2026 estimates land around $10,000 to $30,000 a year, and Radford specifically at $15,000 to $80,000 or more depending on how many surveys and geographies you take.

That opacity is the actual problem. If you run a 60-person US company and someone tells you to "benchmark against survey data," you cannot find out what that costs without entering a sales process, and the sales process is calibrated for buyers ten times your size. This article puts every number we could verify in one place, says clearly which are published prices and which are third-party estimates, and works through when a survey subscription is actually worth buying.

One distinction first, because the words get used interchangeably and they are not the same purchase. A salary survey sells you data. Salary benchmarking software sells you the workflow that turns data into bands you can post and defend. You can buy either, both, or neither. Build a band for a real role below to see what the finished output looks like.

How much does a salary survey cost?

Here is every figure we could verify as of July 2026, with the source type marked. Treat the published column as fact and the estimate column as directional. Anyone quoting you a precise number for a quote-only product is guessing.

The gap between the top and bottom of that table is the entire story of this market. There is a well-served enterprise tier, a well-served free tier that requires you to do the analysis yourself, and historically not much in between for a US company with 60 employees and no compensation analyst.

What actually drives the price

Survey pricing is not really per seat or per employee, even when it is quoted that way. Four variables move the number more than anything else.

Geography. Coverage is sold by country or region. This is the single biggest multiplier, and it is where small buyers overspend most often. Mercer's best membership pricing, a stated discount of over 50 percent, requires buying at least two countries and committing to multiple years. A US-only company buying one year sits at the worst end of that card by design.

Number of surveys. Providers slice data by industry and job family. A technology company that also employs finance, legal and facilities staff often discovers one survey does not cover its whole org chart, and the second and third subscriptions are not discounted the way you would hope.

Participation. Most survey pricing assumes give to get: you submit your own incumbent pay data and receive a participant rate. Mercer offers participants a substantial discount. Not participating usually means paying full freight, so the sticker price you were quoted may assume work you have not agreed to do yet.

Contract term and escalators. Multi-year commitments unlock the good pricing, and comp data contracts commonly carry an annual escalator in the 3 to 7 percent range. A three-year deal at $20,000 is not a $60,000 decision, it is closer to $63,000, and the renewal after it starts from the escalated base.

The costs nobody puts in the quote

Two line items never appear on the order form and both are larger than small buyers expect.

The first is job matching. Survey data is only as accurate as the mapping between your roles and the provider's job codes. Someone has to decide whether your Head of Operations is a functional head or a senior manager, and that one judgment can move a benchmark by tens of thousands of dollars. In a company with a comp analyst this is a skilled task done well. In a company without one it falls to a founder reading a job library for the first time, and a bad match produces confident, precise, wrong numbers. That is worse than no data, because you will defend it.

The second is the submission itself. Preparing a clean incumbent dataset, mapping it, and getting it in before the window closes is a real internal project. Mercer's 2026 participation window has been reported as running roughly March or April through June or July, with results arriving after aggregation. That cadence works for an annual comp cycle. It does not help on the Tuesday afternoon when a candidate asks for the range and expects an answer before the call ends, and increasingly that candidate arrives with their own number and a coached argument for it.

Where can I get salary survey data for free?

The Bureau of Labor Statistics Occupational Employment and Wage Statistics program publishes free US wage percentiles by occupation and metro area, drawn from a federal establishment survey covering millions of jobs. For US employers it is the strongest no-cost source available, and it is the base layer most affordable benchmarking tools build on, including ours.

Its honest limits: it reports occupations rather than your specific job titles, it lags the market because of the collection and publication cycle, it does not cover equity or bonus structure, and it thins out at the executive level and in narrow specialisms. For the roles most companies under 200 people actually hire, an accountant, a controller, a software developer, an operations manager, a sales rep, it is genuinely good data that holds up when someone asks where your number came from. Our guide to how much to pay employees walks through pulling it by hand.

Are salary surveys worth it?

There is a clean way to run this decision, and it does not involve a feature comparison. Work out what a one percent payroll error costs you, and compare it to the subscription.

At 60 employees averaging $85,000, payroll is about $5.1 million and one percent is $51,000. A $20,000 subscription that improves your accuracy by one percent pays for itself twice over, in theory. The catch is that a survey does not improve accuracy on its own. It improves accuracy when someone competent matches the jobs, reads the percentiles correctly, and translates them into bands. Without that person, you have bought precision you cannot use.

So the practical test is two questions. Do you have someone who can do job matching properly, and do you hire in markets or roles that public data covers poorly? Two yeses and a survey subscription is a good purchase. Two noes and you are buying a dataset that will sit unused after the first renewal, which is exactly why plenty of small subscriptions quietly lapse in year two.

Salary survey versus salary benchmarking software

These are different products solving different halves of the problem, and conflating them is how buyers end up with the wrong thing.

Enterprise platforms bundle both, which is why their pricing looks like survey pricing. Lighter tools license or use public data and charge for the workflow instead of the dataset, which is why they can publish a monthly price at all. We take the second approach and list our prices publicly. If you want the vendor-by-vendor version of this, our roundup of the best compensation management software compares ten tools on data source and pricing transparency, and the Mercer alternative page breaks down each Mercer product separately.

How to buy well if you do buy

If you have decided a survey subscription is right, five things reliably reduce what you pay. Ask for participant pricing and actually participate. Buy only the geographies you set pay in, not the global package a rep opens with. Get the annual escalator written down and negotiated, because 3 to 7 percent compounding is real money over a three-year term. Name the alternatives you evaluated during the negotiation, since buyers who anchor against competing quotes consistently land below the opening number. And ask what happens to your access if you skip a participation cycle, because the answer is sometimes a price change you did not expect.

One more thing worth doing before any of that: build a band for one real role from free public data and see how far off it is from what you currently pay. If it is close, you have just answered the question for the price of an afternoon. If it is badly off, you have a much more specific problem to take into a sales call, and a much stronger position in it.

Questions people ask