The Manager's New Brief

11.1

Ch. 11 — Motivation, Belonging, and Psychological Safety

Part III — Leading people

💡 Core idea

People do their best work when they can take interpersonal risks and still see meaning, standards, and a place for themselves on the team. Safety is not permissiveness. It is what makes high standards discussable.

A team that cannot disagree will not catch errors, including errors made faster by tools. A team that cannot fail in the open will stop experimenting. Harmony is not the same as health.

Charity Majors asks whether a happy team means you are doing a good job. Not by itself. Stay SaaSy complements this from the other side: belonging includes knowing the manager will not throw the team overboard in public.

What psychological safety is and is not

It is not the absence of criticism, and it is not a mood you can mandate. It is a practical condition: people can take interpersonal risks — raise a doubt, admit a miss, challenge a plan — without being humiliated or punished. Your reaction to the first piece of bad news this month is the real policy.

That definition follows Edmondson (1999): a shared belief that the team is safe for interpersonal risk taking — not permissiveness, and not cohesiveness, which can suppress dissent. In 51 work teams, psychological safety was associated with learning behavior, and learning behavior mediated the link to performance. Google’s Project Aristotle later found, across 180 Google teams, that how the team worked together mattered more than who was on it, and ranked psychological safety first among five team dynamics. In Google’s later official summary, sales teams that worked together effectively exceeded their targets by an average of 17%; teams that did not fell short by up to 19% (re:Work, 2025).

Safety and standards rise or fall together. Fear can produce compliance. It does not produce the information you need.

Belonging without lowering the bar

Belonging is designed. Watch who speaks, who is interrupted, and whose ideas are credited. New people, quieter people, and people unlike the current majority should be able to see that disagreement is allowed and that credit is fair. If only two voices carry the meeting, do not mistake silence for agreement.

Maarten Dalmijn argues that collaboration needs “troublemakers”: people who will challenge silos without wrecking the relationship. A team that cannot tolerate that kind of pressure will look pleasant and move slowly, or move quickly in a herd.

Motivation lives in the work

You cannot install purpose with a poster. You can design work that has a clear aim (Chapter 5), enough autonomy once alignment exists, and a chance to get better at something that matters. Recognition should be timely and specific, tied to real behavior and outcomes — the same standard as praise in Chapter 9. The 37signals Manager Playbook offers a short template: say thank you or well done, “Name specifically what your report achieved,” and “Explain how you felt about their achievement and how it had a positive impact.” Vague praise is cheap; specific praise teaches the team what good looks like. The same playbook’s Multiplier Managers chapter states the other half: “The behaviors you as leaders reward become aspirational, and what you ignore becomes permissible.”

Suresh Choudhary argues that caring about balance is a reason to step into management, not to avoid it. That is the right frame for this section: load and recovery are management issues, not private lifestyle failures.

Energy, load, and burnout

Look at workload and recovery on a rhythm, not only after someone breaks. Time off that cannot actually be taken is not time off.

On burnout, be careful. Commoncog offers a research-oriented guide; do not casually import clinical language. Lenny’s Newsletter discusses an “ARMOR” frame used by some people in tech. Read the original before you teach that acronym as a method; until then, treat it as a source to open, not as a tool already on your kit. The usable principle does not require a slogan: burnout is a signal about work design, load, and support, and it is yours to notice early.

A Gallup analysis of U.S. workforce panel data (Makridis, 2026) finds that frequent AI users — daily or several times a week — were more than twice as likely as less frequent users to fear their job would be eliminated within five years. Familiarity with the tool did not reassure them. What did move the number was management: workers who strongly agreed they were treated with respect, or that the organization cared about their wellbeing, were about six to seven points less likely to report that concern, and the buffer was largest among frequent users. The data are observational (nearly 30,000 unique worker observations, 2023 through early 2026). Chapter 16 returns to this in change communication.

Tools

🤖 AI-era tip: Do not dress surveillance up as care

Be extremely cautious about always-on sentiment tools, activity monitoring, and “engagement analytics” that people did not meaningfully consent to. Prefer visible, opt-in rituals and conversations. Psychological safety collapses if the team believes the manager’s tools are spying on them. If your company mandates a tool, be honest about what it collects, what you look at, and what you will never use in a performance discussion.

Takeaways

Evidence