Should I give an overview of my research program at the beginning or at the end of my job talk?

Question

I am preparing to give a job talk for an on-campus interview for a tenure-track assistant professor position. In many of the job talks I have attended, candidates will spend 40 minutes talking in depth about a particular project, as well as spending 5-10 minutes on their research program, i.e. summarizing other projects they have completed and briefly describing questions they are interested in exploring in their future research.

Question: Is it better to give the the "research program" part of the talk before the in-depth project, or vice versa?

My personal inclination If I recall correctly, I have seen candidates do either of the two approaches. I am inclined towards talking about the "research program" first, then the in-depth project second. This is because there has been research showing that after 10 minutes, the people attending a lecture lose attention. (See John Medina's book Brain Rules) If I leave the "research program" to the end, I am afraid many people may have "fallen asleep" by then, but if I put it in the beginning, hopefully it is more likely to be remembered.

Answer 1

I agree with your inclination: explaining your research programme, arguably the most important part of a tenure-track job talk, at the end seems counter-productive, to say the least. Moreover, I would definitely not spend only 5 minutes of your talk on your programme. 10 minutes seems like the absolute minimum in a one-hour job talk.

More generally, job talks tend to consist of five parts:

1. The "About Me" part that usually kicks off the talk. Keep this short, as it is essentially pretty boring for the audience.
2. The "Intro" of the research field. Explain your audience what you are actually doing. Here, the point is not so much to sell your own work within the field, but to sell the field itself to the faculty.
3. Your "Research Agenda" within the field. As I mentioned above, arguably the most important part of the talk. The biggest mistake here is to step into what I call the "parameter optimization" trap - making it seem like your research agenda consists primarily of small incremental improvements of your PhD thesis.
4. The "Zoom In". Take out one part of your work, and explain in detail what you have been doing.
5. The "Outlook". Wrap the talk up with an outlook. Where is your field going, where are you going, how would your joining change the research of the faculty?

One problem is knowing how far in-depth each of these segments should be. My PhD advisor used to say that bullets (2) and (3) need to be easy enough to understand that every CS faculty can follow, but deep enough that it still perceived as important and valuable research. For bullet (4), it is ok (maybe expected) that the details are not all that accessible anymore for faculty working in very different fields. However, at Bullet (5), you need to pick up your entire audience again.