Research management: some tips and resources

I’m often asked for tips on managing research projects and researchers – here are some tips, templates, and resources.

Planning a research project

The beginning of every research project can be quite daunting. It’s much less so if you have a plan! The most important parts of a research plan in my opinion include (in a policy context): 

  1.  Path to impact: what impact in the real world do you want to accomplish? Who is your target audience?

  2.  Resources: who will work on the project? What’s your timeline? What methods will you use?

  3.  Audience engagement plan: how will you ensure your research reaches your target audience?

  4.  Failure modes: in which ways might your project go wrong? What can you do now to mitigate these potential issues?

  5.  To-dos: list your next steps as concretely as you can. Include a list of people you will reach out to for feedback later on.

The goal isn’t to have a perfect plan (and you’ll likely stop looking at this plan later into the project), but to get you to scope your project in an organized way and help you have a clearer sense of how to tackle it. As the saying goes, plans are worthless, but planning is indispensable.

Here’s a template I use to get you started: bit.ly/researchplantemplate 

Speedrunning a research idea 

One strategy to get your research off the ground is to do a “speedrun”: a short (usually 5-15h) structured research project aiming to produce a bottom line on a specific question and an output you can share with others. 

The best research outputs are iterative processes: the researcher writes down something, gets feedback from peers, refines their output, and repeats the cycle until they have a polished product (or reaches their deadline). It’s very rare that good research happens in a vacuum with a researcher immediately jotting down brilliant ideas onto blank pages – but working in isolation for too long is a common failure mode among researchers. Try a speedrun to prevent it!

Here’s a workshop I gave for Bay Area researchers on research speedruns in 2024 with some tips and exercises: bit.ly/researchspeedrun 

(Credit to Marie Buhl for preparing the original slides for this workshop.)

Managing a researcher (and being managed)

Most researchers have a supervisor, mentor, or manager. This is a very important relationship that has the potential of accelerating a researcher’s trajectory and multiplying the manager’s impact. But it’s often treated as a bureaucratic, unproductive hassle.

I think this relationship should be treated with the utmost care. For this, I like having some structure, thought, and preparation imprinted on it, and use templates for my weekly and monthly meetings with my managees: bit.ly/weekly1-1 and bit.ly/monthlystepback.
(Many aspects of these templates have been developed by Michael Aird and Peter Wildeford, research managers I’ve had the luck to work with.)

Previous
Previous

Verification mechanisms for international AI treaties

Next
Next

AI Safety Institutes: literature overview