I'm Associate Prof (8 years since PhD) and I've tried to stick to this workload for most of my career (I think this advice came from my advisor Amy Finkelstein, though she probably doesn't follow it anymore since she has a large team of people working with her now):
- 2 projects actively working on (running regressions, writing code, writing/re-writing draft)
- 2 projects revising (to re-submit or submit to a new journal)
- 2 projects in "early stage" (negotiating data agreements, doing preliminary/exploratory analysis)
Given publication lags, this allowed me to stay on track to meet Glenn Ellison's advice to try to publish 1-2 papers a year. When my workload gets higher than this, I've often felt exhausted/inefficient/over-committed. Right now, I have 3-4 papers I'm revising to send back. So as a result, the number of "early stage" projects is being held down because I don't want to start anything new until I get those revisions back on editors' desks.
Even back in grad school I was usually pretty close to this workflow. I found coming up with a "Job Market Paper" pretty stressful, so I found it nice to have other projects to work on to preserve my sanity (I had two revisions to work on my last two years, and two projects that would become the other two chapters of my dissertation).