That the most important thing in any project is to first make sure YOU have understood what the customer/management wants and that THEY have understood what you are going to do.

Seen so many projects (other people's and my own) go down the drain because people had false expectations.

Also, good management is enormously important for any medium-sized endeavour. I cannot count how often I have seen mediocre to bad middle management screw things up, and devs pull off overnighters and 60+hours work weeks for this super-important project that then goes to the bin because "it turned out to be not so important in the end, sorry."

This is the type of stuff you don't (can't?) learn before your first job, and the ONLY way you are going to figure it out is by bumping your head into it.

For most/all of the projects I've worked on, even if everyone on the project is painfully aware of this pitfall, even if we've gone to great lengths to try and understand what we're supposed to be making beforehand, we've still had to take a time out after a couple of months and discuss what it is we're ACTUALLY supposed to be making.

I still remember one of the most useful slaps on the wrist I got from a senior after suggesting how I think this could be done better: "Their expectation is -". Either I'm not understanding the requirement, or I'd be making something they didn't ask for.