How do you wrestle with the importance of being data-driven/tracking progress for community health <> not potentially run the mistake of optimizing for the wrong thing in community ?

Seen this dynamic a few times, curious if anyone has any opinions.

cofounder @DeepALTR (acq. by Commsor).   building better cultures>>better communities>>better outcomes  https://twitter.com/ngaurav38  community x crypto x saas

I never really liked the word "data-driven". I rather always say "data-informed" because at the end of the day, data suggestes something, but I always talked to users to dig deeper and find out what they think and if it's really worth the time to optimize it.

Product Manager & web3 explorer! @purple #56. twitter.com/dbdogan

Ever feel like you're in a situation where the 1on1 sentiment says one thing,but the overall trends says another? 

For ex., w product where a user says they like "x", but analytics data says the opposite. 

Curious w how to deal w something similar but in managing community.

So my experience comes from web2, but it could be applicable to web3 communities,too.
I had a slightly different situation, where I wanted to sunset a product who was stuck in the MVP stage for too long because of other business priorities, but maintaining it was quite annoying.