"A healthy company does $200k of revenue per employee." You have heard this in board meetings, blog posts and VC twitter threads. It is one of the most repeated benchmarks in business, and it is almost always quoted without a denominator. Per employee of what kind of company, exactly?
“$200k revenue per employee is the benchmark for a healthy business.”
Source: commonly cited SaaS/startup rule of thumb; for calibration, SaaS Capital's 2025 survey puts the median for private SaaS at $129,724
Directionally useful, sector-blind in practice. Across 2,573,083 companies with reliable revenue and headcount data, median revenue per employee ranges from $173k (Financial Services) to $329k (Food & Beverage), a 1.9x spread. The $200k line sits below the market median for every asset-heavy sector and above it for most services sectors. Used without a sector adjustment, it will mislead you in both directions.
Source: Veltria dataset, 4,840,019 companies, computed 2026-07-09.
Three things the averages hide
- Capital intensity, not quality, drives the top of the table. Logistics and food production convert assets into revenue with thin margins; their high revenue-per-head says nothing about profitability.
- Technology's median is $198k, closer to SaaS Capital's $129,724 than to the meme. The $200–400k/employee numbers celebrated in public SaaS are the survivors' club. The reality across 251,996 market tech companies is more modest, worth remembering when a target's metrics look "bad".
- Use quartiles, not points. The interquartile range within a single sector is often wider than the gap between sector medians. We ship p25/median/p75 per sector with the dataset.
How to use this
If you are benchmarking a target, a client or your own company: first match the sector, then the revenue band, then compare within the quartile distribution. That requires a denominator dataset, which is precisely what a general-purpose rule of thumb is not.
Benchmark against your exact peer group, filter by sector, country and revenue band.
Get this data →
