How the Monard X Salary Benchmark works
Every report is built fresh when you run it. This page explains where the figures come from, how they are checked, and how much evidence sits behind each one.
Where the figures come from
When you run a benchmark, the tool searches for published pay data for your role in each location you picked, using the same search wording every time, so the same report finds the same sources. It looks for salary guides and surveys published for that country, salary websites, job adverts that state the pay, and contract rates. AI then copies each figure exactly as published, with the sentence it appears in, and a figure is only used if that sentence is on the page.
We do not list the individual sources in your report. What you pay for is the work done on those sources, set out below.
How every figure is checked
Before any figure is used, fixed rules (not AI) turn it into an annual gross salary in the local currency. Monthly figures are multiplied by the number of salary payments that are normal in that country, which is 14 in Spain and Austria.
A figure is thrown out if it is
- net pay (after tax)
- a contractor rate rather than a salary
- in a different currency
- more than two years old
- missing its pay period
- far out of line with national income in that country, using World Bank figures
- less than half or more than double what the other sources say for the same level
How the salary tables are built
The AI builds the salary bands from the figures that pass, and only from those. Every figure that passes is used, including figures published for the levels either side of yours. Fixed checks then run on every table. They catch levels in the wrong order, bands out of proportion, misplaced zeros and tables given per month by mistake. If the median for your level is more than 10% away from the middle of the published figures for your level, every figure is scaled back into line with it. Your report lists any figure these checks corrected, under Data Checks.
How much evidence is behind each location
Each location in your report carries a rating, based on how many published figures passed the checks. The number has a + because the AI also draws on its wider market knowledge to shape the salary bands, so treat it as a minimum. The rating shows how much published pay data was found. It is not a statistical confidence score.
- Strong evidence. 6+ published figures.
- Moderate evidence. 2 to 5 published figures.
- Limited evidence. Fewer than 2 figures passed the checks, so the tables for that location are AI estimates. Treat them as a starting point.
What AI writes and what it does not
The market commentary (pay structure, employers, hiring, cost of living and outlook) is written by AI. Read it as background to the figures.
These parts are calculated by fixed rules, with no AI involved
- the comparison against an advertised salary you enter
- the inflation and unemployment figures, from the World Bank
- exchange rates, from the European Central Bank, or official fixed rates and Exchange Rate API market rates for currencies it does not publish
- the evidence rating
What a benchmark is not
It is an estimate of the market rate for base salary. It is not a quote from any employer, and it does not include bonus, equity or employment costs. Use it alongside your own research.