When you know the theoretical (perfect world) model, it can still be applied in the real world as long as you're aware that you won't get an exact answer, but it will often be "good enough". One of the few times I was able to apply my physics knowledge in my job, I was able to hack together the relevant equations and constants to give realistic answers (to do with still-air cooling of a mobile phone on a charging pad in a car) and I was able to estimate roughly how accurate the answers were, given the remaining unknowns.

I guess I also struggle with knowing when it's ok to be a little vs a lot imprecise, and what that means in terms of units. If I'm building something at kilogram scale can I be off by a few milligrams?