Market share is what people bring up when they have run out of technical objections.
“But Drupal's market share is tiny.” It gets said in a procurement meeting as if it settles something, and I hear it constantly.
That figure counts installs: every abandoned food blog from 2013, every one-page site somebody spun up on shared hosting and then forgot the password to. It is a count of things that exist. It is not a measure of anything you are about to build.
Nobody in that room is building a food blog.
The question nobody asks out loud is narrower: who runs a multilingual, multi-team, WCAG-audited, penetration-tested platform on this, and did they regret it? Different question, very different answer.
The Australian government runs govCMS: more than 300 government websites for over 100 agencies, on one platform. One codebase, one security team, one upgrade path, one audit. US federal agencies, UN bodies, and a long list of universities landed in the same place independently.
None of them picked it from a pie chart. They picked it after people whose entire job is finding reasons to say no had a go at it and failed.
The market-share argument also quietly assumes popularity and suitability are the same measurement. They are not, and everyone in the room knows it; they just do not want to be the one to say so.
So next time someone waves a percentage at you, ask a different question. How many of your sites run on one shared codebase, and what happens to all of them the morning a critical advisory drops?
That number actually tells you something.