Skip to content

Psychological safety

You can build the best project-tracking system in the world. You can run perfect retros, set ideal WIP limits, and run RACI'd meetings. If the people on the team don't feel safe surfacing problems, none of it works.

The research term is psychological safety. Amy Edmondson defined it in 1999 as "a shared belief that the team is safe for interpersonal risk-taking." She published the canonical book in 2018: The Fearless Organization (Edmondson, 2018).

Why it matters more than process

Google ran a multi-year study called Project Aristotle (Duhigg, 2016) trying to find what makes teams successful. The single biggest predictor — bigger than talent, bigger than team composition, bigger than tenure — was psychological safety. Teams where people felt safe to take interpersonal risks outperformed teams that didn't.

In project terms: if people don't feel safe raising concerns about estimates, problems hide until they explode. If people don't feel safe to admit they're behind, status reports become fiction. If people don't feel safe to write honest feedback, the Feedback Register fills with cosmetic comments.

The most expensive thing in your project is the truth that someone knew and didn't tell you.

The four observable signals

Edmondson identified four things people only do on psychologically safe teams:

  1. They ask "stupid" questions — without prefacing them with "this is probably a stupid question, but...".
  2. They admit mistakes — without worrying about consequences.
  3. They challenge the leader's idea — including in front of the leader.
  4. They propose a wild idea — without being mocked or quietly downgraded.

If you've worked on a team where any of these felt risky, you've worked on a low-safety team. Most teams are.

How a Team Lead builds it

You can't give a team psychological safety. You can only do things that make it more likely. The biggest leverages:

  1. Model the behaviour. Admit mistakes publicly. Ask "stupid" questions yourself. Change your mind when someone shows you new information. Do this in meetings, not just one-on-one.
  2. Reward truth-telling. When someone raises a problem, thank them. Even if the problem is uncomfortable. Even if they could have raised it earlier. Especially then.
  3. Don't punish honest estimation misses. If someone estimated 8 hours and it took 24, the conversation is "what did we learn?" — not "why didn't you deliver?" (See Estimation.)
  4. State the Prime Directive in retros. Every time. (See Retrospectives.)
  5. Separate the work from the worker. Critique the brief, not the person who wrote it. Critique the estimate, not the person who proposed it.

How a Solo Operator uses it

This sounds like a team-only topic. It isn't.

The way you talk to yourself about your own mistakes shapes whether you keep showing up to your system. If your inner voice treats every missed estimate as evidence you're bad at this, you'll stop estimating honestly even with yourself. The Feedback Register on your own projects (S-3.2) is most useful when you can be honest in it — and that requires being kind to yourself.

The discipline is the same: what did I learn? — not why am I bad at this?

What this is not

  • Niceness. Psychological safety is not "everyone is nice and never argues." High-safety teams argue more, because they're not afraid to. The arguing is about ideas, not about people.
  • Avoiding hard conversations. Performance conversations still happen. Underperformance still gets addressed. The difference is that people are not punished for telling the truth about what's happening.
  • Consensus. A safe team can still have a leader who makes the final call. Safety means people speak up before that call is made.

Where to read more

  • Edmondson, A. (2018). The Fearless Organization: Creating Psychological Safety in the Workplace for Learning, Innovation, and Growth. Wiley.
  • Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly. The original research paper.
  • Duhigg, C. (2016). What Google Learned From Its Quest to Build the Perfect Team. New York Times Magazine. Project Aristotle.
  • Brené Brown's Dare to Lead (2018) is the popular-audience version of the same ideas.

← Leading vs lagging | Next: Configuration baseline →