How many papers do I need to publish during my PhD?
A simple question with a not-so-simple answer.
I often get the following question from prospective PhD students and first-year students: “How many papers do you expect students to publish during the PhD?”. In some ways, this is a completely reasonable question. Papers are the most tangible outcome of PhD research, and the main yardstick that a researcher is judged by. However, I don't find this question to be a particularly useful way to think. Below, I will try to explain my rationale and encapsulate my advice as a constrained optimization problem.
First, the answer to the question above in some sense is very simple: as many high-quality papers as possible! I don't set an upper bound on how many papers a student can or should publish. When I was starting out as a PhD student, I remember counting 40+ papers that a recent PhD graduate had published during their six-year PhD (approximately half were first-authored, and all were in high-quality venues). There are many similar examples in recent years. I certainly don’t have any expectation that a student will publish as many papers during their time in my group, but I also don’t have an expectation that they won't.
The question could be interpreted to mean: what is the bare minimum number of papers you need to defend your thesis? The answer to this is approximately 2 first-authored papers. But, this is a low bar and not one that should be an explicit goal a priori.
I think a better version of the question is: how many papers do you need to get your dream post-PhD job? The answer to this depends on the job and also, to some lesser degree, on the job market. My main recommendation is to know roughly what this number is right from the beginning of your PhD. In my experience, for academic positions in robotics, 7-8 papers (with at least half first-authored) is a minimum to be seriously considered for faculty positions at the top places. With 10+ papers (including ~2 journal papers for MechE/Aero positions), you are typically above threshold for faculty positions (and then other factors become the deciding ones). For industry positions, there is a lot of variance. For some positions, 4+ papers can be sufficient as long as there is a good fit with the position. For more research-oriented industry positions, 7+ papers (or even more) seem expected.
I recommend studying the CVs of people who have recently graduated and accepted a position you're excited about to get a sense for the expectations. By the time I applied for faculty positions, I had collected a dossier full of CVs (and other materials including research/teaching statements) scraped from websites over the years. This gave me a clear sense for the expectations from the places I was targeting.
Ultimately though, as with any example of Goodhardt’s law, it is important to remember that the number of papers published is an imperfect proxy for your true objectives. You entered the PhD because you were brimming with excitement and curiosity for a research topic — pursuing this curiosity and seeing where it leads is the main goal.
My best overall advice is to pose the following constrained optimization problem for yourself:
In my view, thinking about things in this way helps strike a balance between the primary objective of pursuing curiosity-driven research and the practicalities of getting a job after the PhD.