Yeah, I'm a bit early on this since we are currently in application / interview season, but it's something I have been thinking about a decent amount recently. Why? Because there are things I should have asked for that I didn't and a few things that I did ask for, which have saved my ass. Having done a postdoc outside the country, I wish I had had a bit more information than I went in with, as some of my assumptions turned out to be false.
So, based on my experience and the collective experience of the readers here, I thought we might be able to put together a decent guide for those of you who will hopefully be in the position to negotiate in a few months. The tricky thing is trying to put a number on this, because there is enormous variance between institutions. Instead of talking numbers, I would like to talk about categories of things to ask for. These should be widely applicable, with variable values depending on your field and institution.
Listed based on priority:
This is obviously one of the big ones. You need to have the cash to set up your lab in a way that allows you to get the things done you need to. Everything from major equipment to chemicals to the stupid little spatulas you use to weigh those chemicals out with needs to be thought of. I did it by wandering around my postdoc lab and writing down everything I saw. It is amazing how many little things you can forget / take for granted. List them all.
How you approach putting a number on this category depends on how nit picky the person you are negotiating with is. In my case, I made categories of stuff (e.g. "freezers and refrigerators"), thought about how many I needed and looked up some ballpark prices. In most cases, you are pretty safe going with the list price, because every vendor will crawl out of the woodwork to give you a discount when setting up a new lab. If you use the list price as your guide, you'll have between 20% and 50% of that money left over when you actually by the item. That is a good thing.
In some cases you will be offered the use of shared equipment for big things or possibly that the university will buy you a specific item or has one they will give you. In this case, GET IT IN WRITING. And I mean the exact model number you want and everything. If they say they will buy it, get a timeframe in your offer letter. Even if you are negotiating with the nicest people in the world, get that shit in writing! Decide what you can share and what need to be dedicated to your lab and make sure you get the latter category. There is nothing more stifling than having to wait for shared equipment when you are just getting things to work.
This is another category that you can budget list prices for and come out ahead when vendors battle for your shiny new start-up money, but put the funds in this category to keep you going for longer that you think. I would say at least 3 years is a good idea, 4 if you can get it. Again, gloves to weigh boats, figure out what you need.
There's no use having a stocked lab with no one in it. Figure out support for your people. Does your department offer TA support for students? If so, get a few lines of support committed to your lab. Do you need a technician? Get salary. Same with a postdoc. You may not be able to ask for the moon and the stars, but decide what you want and fight for it. In this funding climate, best to ask for one more year than you think you will need, as well.
And don't forget yourself. If you are in a 9 month appointment, ask them to cover summer salary for a few years. If you get grants to cover it instead, I'm sure they won't mind.
This was one I forgot and have added in response to comments. In most cases, it would be a good idea to add in the costs involved with data production, if you require outside services or expensive methods.
SPACE / RENOVATIONS
This is a category that some people take for granted and I've seen some get burned. It is common for a school to show you a space when you interview. Is that the same space you will inhabit? Maybe, maybe not. Will there be renos done? Probably. But when will that happen? In some cases I hear people complaining about never done customization a year or two into the job. At that point, you've either found a work around or are too set up to allow a work crew in the place to mess it all up. Again, get a time frame in writing!
If you can help it at all, don't teach in year one. There's a lot going on and a million things to figure out already, having to do all that with teaching responsibilities on top of that makes life a lot tougher. Ideally, getting out of teaching in year 1 and then getting some softball grad class in your specialty in your third semester before taking on a undergrad class in semester 4, is not a bad way to ease in.
Do you need to go places to collecting things for your research? Do you want to go to conferences in your first couple of years to get the word out that you've moved and you're a big kid now? Budget it. What about your lab peeps? They'll probably want to go to conferences too, or maybe they need to travel for workshops or collecting. Don't forget to make a category for this!
COMPTERS AND SOFTWARE
This is a funny one, because sometimes money needs to be in a specific category to buy computers. Find out about that. It may be just for grant funds, but worth asking. In any case, you'll need things like computers for you and your peeps. Printers and ink. Back up systems? Servers? What about software you use? A lot of it can be pricey, just look how much the Adobe suite is. And that's before you get to specialized proprietary software.
I'm sure there are some things I am leaving out that others can bring up. It's a daunting process trying to figure out everything you will need to get going, but it can also be a lot of fun.