If first you don't succeed, don't try again.
Maybe.
In some cases.
But not always.
I can think of four cases I've personally had to overcome when trying to impress the importance of accessibility upon developers and failed.
When they think accessibility means it works on my computer
If they think they are the representative sample of their customers, you're not debating standards. You're fighting to explain how they can't be like the 1 billion people affected by one or more disabilities. That's too big a number for anyone to comprehend.
When their defense is their analytics don't show any disabled users
Ah yes, because analytics famously track people who gave up trying to use an inaccessible website. If they're waiting for data from tools that can't even measure the problem, you're not debating accessibility. You're debating logic.
When they argue keyboard navigation ruins their cool designs
If their design philosophy is "Steve Jobs wouldn't have let a screen reader interrupt his keynote," you've already lost. Some people won't believe in accessibility until their fancy dropdown menu traps a VC's CEO in a focus loop.
When they swear they'll do it in the next refactor
The next refactor is like the next book from A Song of Ice and Fire. It's never happening. If they've postponed it three times for that mythical cleaner codebase, let them enjoy their technical debt.
Side note, I don't want to dunk on George R.R. Martin. A Song of Ice and Fire has been an absolute delight to read. I devoured those books. And when I got to the last page of the last one, I just immediately searched online when the next one was coming out. That was over 12 years ago. Still waiting...
Here's the thing.
Of course, you should keep advocating. Maybe try to do it smarter. Switch tactics.
Scare them with legal risk. Bribe them with "Google ranks accessible sites higher" or make accessibility part of their performance reviews if you have the power.
When all else fails, it's better to move on and save your energy for others that are more willing to listen.