If you change what you measure, you change what gets done.
Knowing what to measure is one thing. Getting anyone to act on it is different problem entirely.
Accessibility fixes live in a backlog somewhere, filed under "nice to have." They sit there because nobody has attached a number to them that a product owner (PO) recognises. Support contacts, return visits, Net Promoter Score (NPS) by access method are those numbers a PO understands.
But the other thing is ownership.
Accessibility falls partly because nobody owns it. It gets handed to whoever raises their hand or whoever complained last. When you measure it like a product metric, it needs an owner. That's too uncomfortable for a lot of teams.
And there's a timing problem.
Accessibility compliance gets checked at the end, before you ship. That's when changing anything is expensive and nobody wants to hear that. So it's back to the backlog for those issues. There's always next time, right?!
If you're measuring the right things, then you can set up tests that you can run early. And early findings are cheap findings. These findings deserve a meeting. What gets a meeting gets a budget.
And what gets a budget gets done.