Contents
180 teams, more than 200 interviews, over 250 attributes analyzed: that is how Google set out to discover what makes a team effective. The company that measures almost everything expected the answer to lie in team composition – the right people, the right mix. The answer turned out differently. And at the top of the list it put a term that until then had been at home mainly in academic research. An example of the Team Principle of Safety.
The Search for the Formula
Situation
The internal research project was called “Project Aristotle,” after the idea attributed to Aristotle that the whole is greater than the sum of its parts. The researchers were part of the HR function; Julia Rozovsky is named as the project lead. The study covered 115 engineering project teams and 65 sales teams, deliberately mixing high- and low-performing teams. It ran for about two years. Google distinguished “teams,” with a high degree of interdependence, from mere “work groups.” Effectiveness was measured in four ways: through the assessments of senior executives, of team leads and of team members, and through sales performance against the quarterly target.
Connection to the team principle
The study design shows how seriously Google took the question. Google didn’t want to confirm what people already believed; it wanted to find out what makes the difference. That is precisely why the result carries weight: it was not what they had been looking for.
What Didn’t Matter
Situation
Many obvious explanations didn’t hold up. No meaningful effect showed up for co-location, consensus decision-making, members’ extraversion, individual performance, workload, seniority, team size or tenure. Google’s central finding: who is on a team matters less than how the members work together, structure their work and view their contributions.
Connection to the team principle
For you as a leader, this is uncomfortable news and good news at the same time. You can’t fix a weak team through better staffing alone. But you can work on how people treat each other – and that is far more within your control than the org chart.
Safety Comes First
Situation
Google identified five factors, in this order: psychological safety, dependability, structure and clarity, meaning, and impact. For the first, Google adopted Amy C. Edmondson’s concept: the shared belief within a team that you can take an interpersonal risk – ask a question, admit a mistake, voice a half-formed idea – without being embarrassed or punished.
Connection to the team principle
This is the principle in its purest form. In 1999, in a study of 51 work teams at a furniture manufacturer, Amy C. Edmondson had shown that psychological safety is linked to learning behavior, and that learning behavior mediates the relationship between psychological safety and team performance. Google’s ranking fits: safety is the foundation that dependability and clarity need in order to take effect. If you can’t ask, you can’t clarify your role. If you can’t admit a mistake, you don’t become dependable – just invisible.
From Finding to Exercise
Situation
The study gave rise to a tool, the “gTeams” exercise: a ten-minute quick check on the five factors, with a results report and a team conversation afterward. According to Google’s own figures, more than 3,000 employees in about 300 teams used it; reported improvements were about 6% in psychological safety and 10% in structure and clarity.
Connection to the team principle
Safety doesn’t come from a lecture about safety. It comes from a team regularly talking about how its members treat each other. The quick check is only the occasion. What really counts is the conversation afterward.
What the Example Doesn’t Prove
Situation
Project Aristotle is an in-house study. It was not published in a peer-reviewed journal; outsiders cannot verify its data and methods. The percentages for the exercise are self-reported.
Connection to the team principle
Independent research also calls for nuance. A 2017 meta-analysis by M. Lance Frazier, Stav Fainshmidt, Ryan L. Klinger, Amir Pezeshkan and Veselina Vracheva, covering 136 samples with more than 22,000 people, shows a link between psychological safety and performance – but how strong it is depends on study design and national culture. And Hong Deng, Kwok Leung, Catherine K. Lam and Xu Huang found that a climate of safety does encourage learning and speaking up, but can also lower work motivation. So the “only” in the principle’s core statement is a deliberate overstatement. Safety is the foundation, not the whole house. Without high standards, it breeds complacency.
Key takeaways
Key Lessons
- Stop looking for the solution in team composition first. Ask how your team talks to one another before you ask who is missing.
- Put one simple test question to your team: Can someone here ask a question or admit a mistake without it hurting them?
- Take ten minutes for a quick check on the five factors – and at least three times as long for the conversation about the results.
- Combine safety with high standards. Say clearly what you expect, and at the same time make it risk-free to report problems along the way.
Summary
Google expected to find what makes teams good in who was on them – and found it in how they work together. Of five factors, psychological safety came first: the best teams were those whose members could ask questions, admit mistakes and share ideas without fear. The study is internal and cannot be verified, but it is consistent with independent research. It supports the Team Principle of Safety: a team only learns when no one has to fear speaking up about a mistake or a doubt.