Feature creep usually starts looking reasonable

i keep noticing feature creep rarely shows up as a dramatic bad decision. it usually arrives as one extra button, one extra setting, one extra workflow that someone swears is necessary.

then six months later nobody can explain the simple version anymore. where do you all draw the line between useful flexibility and just making the thing harder to use?

“One extra setting” is how you end up with a settings page that needs its own search bar. I try to draw the line at “can a new person succeed without touching this? ”—if the answer’s no, it probably shouldn’t be a setting, it should be the default behavior.

I draw the line at “can we delete it later? ” If the new button/setting doesn’t come with an owner and a date where we re-justify it, I’m not interested. In my world the stuff that kills you isn’t the extra UI, it’s the extra failure mode nobody remembers until 3am.