The Next 90 Days

Managing Research Software Projects

Greg Wilson

2026

Learning Objectives

  • Write a 90-day plan with three concrete priorities.
  • Know what to read next.

Planning at the Right Timescale

  • Ninety days is long enough to finish something and short enough to be concrete.
  • Three kinds of work must appear in the plan.
  • Development of new features.
  • Maintenance: bug fixes, reviews, dependency updates, and compliance.
  • Getting support for future work: impact summaries and grant proposals.
  • Maintenance consumes all time if not explicitly allocated.

The Monday Morning Test

  • A plan with five priorities cannot be finished in ninety days.
  • Write a plan with three priorities, each with a clear definition of done.
  • Ask “what is the first thing you will do on Monday morning?”
  • If you cannot answer, the goal is too vague.
  • “Work on the documentation” is not well defined; a specific first step is.

Draft Your 90-Day Plan (12 min)

Prompt the LLM like this:

Based on this project health audit [paste rubric from Session 1] and these open issues [paste items from your project], create a 90-day plan with three concrete priorities. For each, explain what will be done, how I will know it’s done, and what could prevent it from being completed.

  1. Note any priority the LLM suggested that you know won’t happen because of a constraint it doesn’t know, such as a funding gap, lack of team capacity, or an external deadline.
  2. Make a list of two people you need to talk to (either by name or by role) and what you need.

Monday Morning Test (8 min)

In pairs:

  1. Read one priority from your plan to your partner.
  2. Tell them what the first thing you’ll do on Monday morning is.
  3. Ask them for feedback on whether your description is specific enough.

Wrap-up (5 min)

  1. What is the single most useful thing you learned or produced today?
  2. What was included that you don’t think you’ll use?