Recruiting. I wish I knew more about how my department admits students before I arrived. Ask now! In my department, students arrive in August uncommitted to any particular lab. They then pick an advisor in the next 2-3 weeks (i.e. before classes start, so they can pick appropriate classes). It is a huge free-for-all. Professors can request to interview students, and then must sign off before a student can officially join a lab. This is partially because the coursework requirements are really flexible, and can be tailored to meet a student's interests, but means the students must pick a group quickly. The system ended up working for me, since I got 2 great students right away. If students were admitted to a particular group from the start, I would have potentially waited a whole year before having students in my lab.
What I should have done differently, if I knew this in advance:
1. Actually spend time on my group website. This is a great recruiting tool! Students look at the departmental website before they arrive on campus to plan out their potential advisors. Many arrive pretty sure of who they want to work with. I spent valuable time working on my Fall class that I should have spent on this.
2. Make a list of the classes I want my students to take. The flexibility is great, and I really appreciate it now that my students are taking the classes that best fit their projects, but I really needed to have a list available for when I was interviewing students.
3. Really consider what I want/need in a student. I kind of recruited on the fly and got lucky. This year, I will think about my projects, and what kinds of students would make the fastest progress on them. I will actively look at the incoming students' files ahead of time, not just after I set up an interview with them. I feel like I might have missed a few, since I do kind of hybrid research and cover several sub-disciplines in my department.
Teaching. I worked really hard on my class in August. I was teaching a ~200 person class of non-majors. I tried to get way ahead in my lectures and assignments. I spent some time getting WebCT assessments set up so I would be able to push students to keep up without collecting homework. All of this was a waste of time (except for the WebCT assessments, which I will now use every time I teach the class). I had one TA to run the tutorial session (3 hours, one night a week). WebCT sucks by the way, in case you didn't know.
What I should have done differently:
1. Not worry about being a month ahead of the students. I ended up having to redo almost all of my lecture slides after reality hit my lesson plans. I ended up spending the night before class prepping every lecture, anyway, so I may as well have just done that right from the start. The pre-prep was a waste of time.
2. Started writing my exams earlier. Writing exams is really difficult and time-consuming. Considered how long it would take me to grade each problem when writing them. Next time, I will have some long-problems that are multiple choice as well. Partial credit is over-rated! This would also significantly cut down the number of arguments/discussions/beggars for points in my office. No one argues with the Scantron.
3. Ask for more TA support! If I can't get another TA to help grade, I will consider having my TA grade while I do the tutorial. Grading is really, really time consuming, and midterms here coincided with some grant deadlines for a majorly awful few weeks.
4. Do not make appointments with students who don't have a really good reason they can't make office hours. Otherwise, everyone wants one. I held open office hours the day of exams, though, and this was very successful. A surprisingly small number of students show up, but everyone feels like I care on evaluations.
Research. Manpower (other than myself) was the most important thing I could have. Teaching, writing, meetings and service (even the light service done by new recruits) eat up a lot of time. After I trained my two students, I rarely actually spent time on the instruments myself. It is far more important as a professor to keep up in the literature (to keep current and to spark new proposal ideas) than to collect data. This was a big adjustment to me from the National Lab environment, where I was still pretty active in data generation.
I also found that even in academia, there is spare equipment lying around. Many people are happy to give/lend it to a newbie (especially if said newbie will maintain it or give it space). Ask around! I learned this the hard way after spending some of my precious startup on things I ended up getting for free later on. Don't be afraid to ask about what stuff may be available.
Time Management. Everyone will want you to give seminars in their department/guest lectures in their classes because you are new blood. This seems like a minor request at the time, but ends up taking more time than you think. I ended up saying yes to everyone, which is good from a local networking point of view, but bad from a getting things done point of view. This year, I will start to say no more, since I have more things on my plate. Job searches and recruitment eat up a TON of time. You can't get out of this, since it is one of the most important service obligations, and it falls on the whole department. It is very strange seeing a search from the other side on the first year on the TT!