How do you evaluate whether users actually understand better, rather than just feel like they do?

I don't. Too new and I haven't fully committed to this idea yet.

For what it's worth it worked for me.

Nice