How we collect and calculate the salary data

We collect job postings from company career pages and the public job boards of applicant tracking systems such as Greenhouse, Lever, Ashby, Workday and SmartRecruiters. We don't use job aggregators or salary surveys. The current data covers 1,332,319 postings, of which 132,956 publish a pay range.

• We read the pay range from the posting and take its midpoint.
• Hourly pay is multiplied by 2,080 hours and monthly pay by 12 to get yearly pay.
• Other currencies are converted to euros at approximate exchange rates.
• We classify each job title into a profession and role, and read the level, country, work format, company size and industry from the posting and the employer.
• Duplicates are removed: the same job posted in several places counts once.

• We show medians and percentiles, not averages, so a few very high or very low postings don't distort the picture.
• A figure appears only when at least 20 postings back it; a page needs at least 50.
• Country comparisons “for the same job” compare pay for the same role and level against the EU average.
• Skill premiums compare postings that ask for a skill with similar postings without it, with the same role, level and country.

Advertised pay isn't the same as what people earn: it leaves out bonuses, equity and benefits, and final offers can land anywhere in the range. Employers that publish ranges aren't a perfect sample either: in Europe fewer postings show pay, so European figures rest on fewer data points. The data is strongest for the United States, the UK and the EU.

The data was last updated in October 2026.