Inviting speakers

Not talking about conferences here--there are a whole other set of considerations when organizing sessions for a conference. I am talking about invited seminars. The sort of thing that has a limited budget with many slots to fill, including slots at times that are very unpopular for travel. How do you decide who to recommend as a speaker? Do you suggest friends? Big name speakers? People who you want to hear? People who you want to meet?

In my department, there is a seminar coordinator (a service position usually given to newer folks to help them network) who organizes the seminar schedule. This person solicits suggestions from the department for speakers, but has discretion over who to invite. The person who makes the initial suggestion acts as host, though the seminar coordinator takes care of the invitation, scheduling, and travel details. The host organizes the visit schedule, introduces the speaker, and arranges for dinner. So the host gets a lot of contact with their suggested invitee. Thus, I tend to suggest people who I want to meet and people whose research I want to hear more about after seeing a short conference talk.

My suggestions tend towards the early- and mid- career side, as there are many other mechanisms (and prestigious named lectures with actual budgets) that bring in well established big names. I figure that I am more likely to make a possibly useful connection with someone earlier in their career, especially since the seminar coordinator gets to do all the inviting and off campus interacting. Plus invitations are a whole lot more meaningful to less established people. I am sure the big name folks could probably travel every day of the year if they wanted to. Even better, newbies tend to have fewer schedule constraints and are often happy to take slots at less desirable (like January in a place with winter) travel times, since they do less travel overall. Lately, I have also tried to include people I've seen give great talks at meetings who are also visible members of underrepresented groups, even if I don't have a huge amount of research overlap, because I think it is really important to put a diverse slate of speakers in front of our students. I think I saw this idea a few years ago on Drugmonkey's blog (I'll admit that I am too lazy and too much in the middle of F*cktober to go looking, but I think it was there).

I guess I put a lot of thought into something that nets me an average of one hosting opportunity per year. One of my colleagues thinks I am nuts, and only suggests people who are either mega-big names or people directly in his research area. But how hard is it to start a list at a meeting, and just keep adding to it as you see people who might make good seminar speakers in the future?