This is a tip that has taken me years to really grasp, and I still haven't fully internalized it. I don't like to work at the last minute. If I have something to do and I don't do it, then it stays in the back of my mind until I do it. So, especially if it's some boring task like administrative work, I like to get rid of it as soon as possible. That's a mistake. I'm speaking of my experience in France, so maybe it doesn't apply so much elsewhere. The reason it's a mistake is that what you are required to do changes all the time, so the latest you do it, the least work you will have to do.
Every new politician seems to want to add a new layer of bureaucracy, independently of their political origin, so the amount of administrative work you are required to do as a scientist keeps growing, and it doesn't seem to converge. But setting up new rules and reglementations in a complex bureaucratic monster is not easy, so the monster often outputs nonsensical forms and requirements. One example in France is the evaluation of labs (HCERES), whose role is unclear and changing. The amount of redundancy and the absurdity of some requirements is abysmal. For example, you are required to fill a SWOT diagram, to select and list 20 % of all your « outputs », but also to list each one of them in another form, etc. Because many of the requirements are vague and nonsensical, any organization that deals with them will take some time to converge to a clear set of rules issued to the labs. I have written my evaluation document about 4 times because of the changing instructions.
Another recent example is the new evaluation system set up by INSERM (national medical research institution). Someone there (external consulting company ?) apparently decided that having an online CV with fields to fill instead of uploading a text would be more convenient. So for example you have to insert, one by one in web forms, the list of all journals for which you have reviewed in your entire career, and how many papers you have reviewed for each of them. You need to insert the list of all students you have supervised in you entire carrier, with names and exact dates, etc, all one by one in separate fields. Imagine that for senior PIs. Guess what : one week before deadline, the requirement of filling that CV was lifted for most scientists because of many complaints (a bit too late for most of them, including me). About a quarter of them still have to, but the message says that the format of the CV will change next year since it was not good, so all the work will basically be for nothing.
So here is my conclusion and tip : bureaucracy is nonsense and don't assume otherwise ; just set yourself some time on the deadline to do the required work, whatever it might become at that time (and it might disappear).