> Why not collect all the information you possibly can, whether or not you immediately see obvious value to it, and then find ways to justify it later?

Because it's non-trivial and it costs a lot of money to do that, in terms of engineers' salaries, storage, etc.