Office Politics

Managing Research Software Projects

Greg Wilson

2026

Learning Objectives

  • Recognize that every organization is political and that “we don’t do politics here” is itself a political position.
  • Identify who holds meaningful power over a decision by asking whose support the decision-maker actually needs.
  • Explain which “political” skills predict job success.

Politics Is Not Optional

  • Politics is “who gets what, when, how”.
  • It has no implication that the process is fair or democratic.
  • The alternative to politics is coercion, not harmony.
  • “We don’t do politics here” usually means someone has already won.

Selectorate Theory

  • Leaders need enough support to stay in power.
  • They distribute benefits to a minimum necessary winning coalition.
  • Everyone else is the selectorate: large but not powerful.
  • Funded features and approved hires reflect someone’s winning coalition.
  • Identifying that coalition is more useful than making a better technical argument.

Acting on This

  • Political skills predict career advancement more than technical competence.
  • Advocating for yourself and seeking allies is ordinary, healthy behavior.
  • So is timing proposals for when decision-makers are receptive.
  • Explain issues in terms decision-makers care about (which usually isn’t code quality).
    • Refusing to do this is a practical error while feeling virtuous.

What about AI?

  • LLMs make it easy to produce volumes of professional-sounding position papers
  • LLM text degrades the signals of the author’s actual state

Map the Coalition (10 min)

Think of a decision that was made in your project or institution that went differently than you expected based on the technical merits. What are three non-technical factors that might explain the actual outcome, and what information would help you test each explanation?

Rewrite the Argument (10 min)

Take a technical proposal you have made or plan to make. Rewrite the core argument for an audience that cares about something other than technical correctness. Do not use any technical terms your audience would not know. You must frame the proposal in terms of something they already care about, such as risk, cost, reputation, or timeline. Trade your proposal with a partner and exchange critiques.