Chad just posted a bit of pre-tenure advice, including the very important advice to take all advice with a grain of salt. I would say that also applies to the rest of his advice, because I'm about to post contradictory advice. You should also take my advice with a grain of salt. Be aware that it comes from somebody who has been beaten into being very cynical about the system. On the other hand, you can learn from my mistakes.
My advice here is specifically for faculty at a research University, most specifically Vanderbilt. It's primarily for physics and astronomy (indeed, primarily the latter), but will apply to a lesser extent to anybody in the physical sciences. I would hope that the two new hires in astronomy at Vanderbilt will at least read and think about this, even if they decide thereafter that I'm full of it.
Keep your eyes on the prize. Your goal as junior faculty is to get tenure. Your goal is not to do your job well. Of course, those two goals are not orthogonal... but neither are they parallel. Time spent improving how good a professor you are is time wasted if it does not also improve your tenure portfolio. This is extremely frustrating, because of course most of us have some sort of sense of personal pride and responsibility, and most of us care about being a professor. However, it's very important to keep your eyes on the prize, or you might get bitten. The most important aspect of this is:
Don't waste time trying to be a good teacher. Research universities, even the ones that claim to care about that, don't really care about that. Vanderbilt cares about looking like it has good teaching, not about actually teaching well. Your teaching will be judged not by how much the students learned, or even any sort of imperfect assessment technique designed with that goal in mind. Your teaching "quality" will be judged by student evaluations. As such, your goal should be to maximize student evaluations with a minimum of effort. The student evaluations used by Vanderbilt don't, at least in the sciences, correlate terribly well with how much the students learned. If you have bad evaluations, then you are almost certainly a bad teacher. But you can get medium to good evaluations with a wide range of student learning outcomes.
One corollary of this is that you should pay no attention to education researchers like Slater and Prather, and by no means attend the APS/AAS/AAPT junior faculty workshop. After you have tenure, then you can decide if you want to spend time on that kind of thing. All that paying attention to those guys and attending that workshop will do is fire you up with enthusiasm about teaching as an intellectual activity, it will inspire you to do a better job of actually imparting some lasting knowledge to your students, and it will get you excited about and interested in teaching. It will get excited about a whole bunch of stuff that is not in your interest. None of this will be rewarded when you are judged by your university. The single lesson that you need from that workshop I will tell you right now: when you are teaching, you need to tune your assessment to your course goals. If your course goals are for students to understand processes, but you test them on memorization of facts and equations, then what the students will learn is memorization of facts and equations. You are assessed by your student evaluations, so all you should be concerned with is your student evaluations.
It is entirely possible to get good student evaluations by teaching an "active learning" course that uses all of the best practices and materials that have been coming out of physics and astronomy education research. However, there are ways to get good student evaluations that require less investment of time, and are less hazardous. A lot (though not all) have found that when they first start using "active learning" techniques, their student evaluations take a temporary dip. The consensus seems to be the reason is that you are challenging the student assumptions about what a science course is "supposed to" be. Perhaps this is worth it with advanced courses, but when you are teaching a non-majors course (which, honestly, at least half the students are taking for no reason other than to satisfy a requirement), it is much in your interest to avoid challenging those assumptions. Students come in expecting that in lecture, they will just need to passively listen and take notes. They expect that spending a lot of time rereading the book, memorizing terms and definitions, and memorizing equations and the occasional procedure for soling a problem, is everything that is needed to score perfectly in the class. Do not challenge this assumption, because it's not worth it. Teach a fact, definition, and problem method memorization class, because it is much easier and faster for you to do that than it is for you to teach a creative, up-to-date, better, active-learning sort of class. Make sure your lectures are informative, entertaining, and clear, and make sure you are very clear about what it is you are expecting the students to do. And, leave it at that. Extra time spent figuring out how to teach a better class while maintaining your student evaluations is time and energy wasted, time and energy better spent on other things.
Or, really, not "better" spent, for I still am idealistic enough to think that we should be teaching as good of classes as possible, that we ought to be trying to help educate people about how science works. But, alas, that attitude isn't the attitude that is best rewarded at a research university. Remember to keep your eyes on the prize. You aren't judged on that stuff, you're judged on student evaluations, and your goal is not to do your job well, your goal is to get tenure. So while time and energy won't be "better" spent on things other than paying attention to current educational research, that time and energy will be spent much more in your interest.
Obsess about getting funding. Obviously, I highlight this, because it has been my Achilles heel. However, I think it's something worth considering, for a couple of reasons. First of all, and perhaps most importantly, this is the thing that is least under your control. Making progress in your research and publishing is almost entirely under your control. Yes, there are vagaries of weather and telescope time, there are vagaries of the science not working out the way you thought it would, but hard work will lead to publications, and you can just do it. Your classes are under your control. Your funding.... Well, nearly everybody I've talked to about my situation has said something to the effect that NSF astronomy is so oversubscribed, that it's a crap shoot who gets the funding, etc. You can write a good proposal, but who is reading it is just as much (if not more) important than what you write.
You can make sure everything else works, but this one you have to worry about. Obsess about it until you have it in the can. Some specific pieces of advice:
Universities like "continuous funding." Even if you get a three-year grant, keep submitting proposals until you have a second one. If you wait until your grant is one year from expiring, you may be facing that 1-in-5 chance (or worse) of getting an NSF grant, and will be facing a gap (that may never end) in your funding. Ideally, you have grants that overlap, and you have a grant that is active at the moment when you go forward for tenure.
Submit multiple grants each year. NSF astronomy has a call once a year. Make sure you're submitting at least three grants each year. One can be yours, and the others should be with Co-PIs, either from your institution or from other institutions.
If you're an observer, get a Co-PI from an institution that has guaranteed access to a large telescope. Without that, you are at a disadvantage in the eyes of many NSF astronomy panels. (Probably because those panels include people from places that have guaranteed access, and have argued that the limited funding resources are best spent when they go to the places that have the observing resources.)
Figure out who is best funded at your institution and in your subfield, and try to pick their brain as to how they do it. I can give you the name of the astronomer at Vanderbilt, but if you're there it's pretty obvious who it is.
The second reason you need to obsess about getting funding is that it is the skill you were least selected for. True, you probably don't have a lot of teaching experience before starting your junior faculty job, but you have demonstrated an ability to communicate well. If your colloquium was an incomprehensible disaster, you wouldn't have been hired. Lean on those skills for teaching (once again, don't mess around with learning new "active learning" techniques), and lean on your research skills for publishing. You've already got both, and you were selected for that. You weren't selected for your ability to get funding. Given that this isn't something you've been trained to do, and given that funding is so amazingly tight in astronomy, and finally given that funding is an absolutely non-negotiable requirement for tenure at a research unviersity, you must, absolutely must, obsess about it.
Don't worry too much about publishing. Do it, but don't worry about it. The fact is, you have already demonstrated that you're very good at getting research done and published. If you hadn't, you wouldn't be in this job. There are so many post-docs vying for the rare faculty position that the faculty of your institution wouldn't have hired you if you hadn't already demonstrated a good publication record. Keep doing what you're doing, make sure to get papers out. Do collaborate with other people, because then you will have your name on other papers and your publication list will get longer. If you must have a metric, in astronomy one first-author paper a year plus three or four other co-author papers a year should be just fine. A paper first-authored by one of your graduate students should be as good as a first-author paper for you. You can have fewer first author papers if they are highly cited, high-impact papers, but the total number (including co-authored) of papers shouldn't fall below the 3-per-year rate.
My biggest mistakes were caring too much about the teaching, putting a lot of energy and thought into how to teach better, and not obsessing enough about funding. A lot of the time I spent going to teaching workshops and thinking about how to improve either the form or the content of my classes should have been spent working on additional (ideally collaborative) funding proposals. I know I give good public outreach lectures, and as such I know I can give good "present the facts" class lectures. If I hadn't tried so hard to make the class better, if I hadn't challenged students to learn to understand the science in the introductory class instead of just to remember it, I would have fought less of an uphill battle in getting student evaluations (which, by the way, are just fine), and more importantly, I would have spent a lot less time on my classes. It was time wasted for me, clearly, since my tenure case is in the toilet for funding reasons.
The goal is not to be a good professor, the goal is to look like a good professor to those who will judge you. The goal is to get tenure.
Keep your eyes on the prize.