Salary benchmarking

Where benchmark data comes from

Four tiers, in descending reliability: participating compensation surveys (matched role definitions, employer-reported, the professional standard), published government and industry data, the growing pool of posted ranges that transparency laws have created, and self-reported salary sites (directional at best, skewed by who reports). Good benchmarking triangulates at least two tiers and records which market it is pricing against: industry, company size and location change the answer materially.

Matching roles, not titles

The craft is in the match: a "product manager" spans a junior coordinator to a mini-GM depending on the company, and benchmarking titles instead of role content produces confident nonsense. Match on scope, level and accountabilities, benchmark the band midpoint rather than chasing every data point, and refresh annually, or faster in running-hot markets, because benchmark decay is how organisations end up shocked by their own offer-decline rate.

The signals your own funnel sends

Your recruiting data is a free benchmark: offer-decline reasons, the gap between advertised range and accepted offers, and counter-offer sizes all price the market in real time. A rising decline rate on pay is the market repricing your bands for you; benchmarking is just the cheaper way to hear the same message earlier.