re: Jonathan’s #1 — two added bonuses here are that:
(a) the process of asking for more materials as you go along will inevitably be documented at the job search wiki, so even if you’re not allowed to tell candidates when they’re out of the running, they’ll likely know when they haven’t made the next round. So it’s a more humane way to run a search; and
(b) asking candidates for more info gives you a better sense of them as people. I’ve often found that the little back-and-forth emails that accompany this process can be friendly and informative.

My other key bit of advice: HURRY UP. There’s a humane argument and a strategic one here. The humane argument is that academia is ludicrously slow in its hiring, and those of us lucky to have jobs and be on committees owe it to those applying not to leave them hanging for months. If you know you won’t be able to look at the materials till, for instance, Nov 1, don’t ask for them on Sept 1! The strategic argument is that the faster you go, the more likely you will beat other institutions to the punch. I’ve found that as much as, yes, the market is flooded, there are definitely much better candidates and much worse ones out there, so a little speed can help ensure that the better ones are locked up and signed before the others even wake up, take the coffee pot off, and look at the first CV.

And a final one is to be nice. When I hear horror stories about programs and depts from friends who were treated poorly there, I smell mean dysfunctionality, and I won’t recommend friends apply there in the future, I’ll warn prospective grad students away from applying there, and it makes me wonder about the folk working there. Niceness is karmic, in other words, and if you’re not nice — even to the candidates you don’t like — it’ll bite you on the ass.

– I’d recommend institutions consider using Interfolio or a similar online provider to manage the search. We did this last year, and things worked well, especially allowing people to access search materials remotely, even from abroad. It saves hassle for everyone, plus is environmentally & economically efficient.

– I do think there’s great value in some one-on-one sit-downs with candidates, but it’s important for the faculty to remember that the candidate is doing many of these. Before the interview, it could be useful for the committee to divvy up some of the issues that each will address in their one-on-one so it’s not endless repetition of the same material.

– Be honest with the strengths and weaknesses of your department, institution & location. If the search is successful, this person will be your colleague for (hopefully) many years, so a bait & switch will bite you in your ass. A candidate who knows what type of place they’ll be working is set-up to succeed far more than someone who is disappointed to discover that things aren’t what they had expected.

– Try to shelter the candidate from the depths of your bureaucracy & rules. Every institutions has its own quirks, and you shouldn’t subject the candidate to your gripes with how things work. That can mean being clear about expected reimbursement timelines & procedures (if the rules don’t allow the institution to pay for everything in advance), avoiding complaining about limits on dinners, alcohol, etc., and just trying to show candidates that they would be working for a functional institution. Don’t lie, but show them how things work best by successful planning & clarity.

– Be as transparent as possible. I blogged about one search I led, and think it was quite successful to embrace that level of transparency. You might not be able to convince your administration to do that, but in general, make sure you know the legal & ethical limits of what you can share – and then share everything you can within those limits.