When I started my position, I saw lots of advice about balance, teaching, recruiting, setting up, etc, but not much about how to best use startup funds. Startup funds are magic money--not tied to any particular project or any particular budget, they can be spent on anything. This type of money is so amazingly rare in academia that I am unlikely to ever have significant amounts of it again. Watching funds disappear from the account can be painful, since they are in effect irreplaceable.
When I started at Prodigal U, all I had was my startup money to buy equipment and supplies, to pay salaries, to travel, and to cover user fees for instrument access. The temptation (at least for a cheapskate like me) is to horde it--what if nothing else comes in? My new chair and my PhD advsior both told me that it is more common for new profs to cheap out on buying stuff than to burn the startup too quickly, and that you can always ask the department to help you pay for students, but you can never get wasted research time back, so I decided to spend. I took a "burn the ships" approach to my startup. I would turbocharge my first year by spending on anything that would get me going faster and worry about the funding situation later.
This was a deeply considered decision. To be at the cutting edge in my research area requires some very expensive equipment. I decided to buy most of what I will need rather than depend on instrument access in other labs. I was thinking 1) I have priority usage, so we can work 24 hours a day if necessary, 2) we can modify the equipment as needed and make permanent installations of commonly used setups, and 3) equipment money is hard to get and user fees can really add up. That said, burning through my startup like this means I cannot hire a postdoc or technician until I get some decently funded proposals.
So far, it has been a good decision. My equipment is humming along and producing nice data. I had 3 proposals funded this year (1 new investigator, 1 solo small grant, and one as co-PI with 2 other profs). None are huge money, but I can probably support 4 grad students with what I have in hand. I have 2 grad students plus myself as personnel (augmented by 3 summer students), and I will probably take 2 more grad students if I can find the right ones. And then I need to find more money.
The postdoc thing weighs on my mind a bit--would I be going faster if I had one? In my field, postdocs last 1-3 years, so it just didn't seem worth it to me to invest so much time into someone so temporary in the very beginning. Mine is a student-heavy strategy that is only possible because my first 2 students are doing amazingly well in terms of research productivity. Sometimes I wish I had hired a postdoc right away, especially when things are not going well in the lab and there is troubleshooting to be done. But then I see some colleagues waiting around for instrument time while we are using our own, and I am happy again.