Here’s how I do it since we’re sharing:
I do something similar to you but somewhat differently too. I of course have my threat model and I set different thresholds for the tool with respect to its privacy, security, and convenience. This is my three body problem I deal with. And I don’t change my mind often unless and until one of them crosses the threshold to becoming a deal breaker. I think this is relatively a simple way to go about it. So, once one has their own thresholds that you cannot allow a company/tool/product to do something, you can easily decide what to do and what not to. Of course, this only works if you have relatively equivalent alternatives of said tool/product and luckily the privacy space is growing with these.
For example: after years of FF screwing its community with poor decisions and no engagement from them in any real way like how Proton does, the TOS change from them was the last straw for me and I finally moved away.