Psychological Safety Is Not a Feel-Good Factor.

It is a precondition for performance.

When I talk to executive teams about psychological safety, I often hear the same half-sentence: "Important topic, but right now we need to focus on..." — followed by margins, restructuring, the AI roadmap, cost discipline.

I understand it. And I disagree with it.

What many treat as a soft cultural question is in fact a hard performance condition. That isn't my opinion. It's what twenty-five years of research consistently shows.

What Edmondson actually found

Amy Edmondson, now a professor at Harvard Business School, published a study in 1999 that effectively founded a research field. She examined 51 work teams in a manufacturing company — and found something counterintuitive: the better teams reported more errors, not fewer.

Not because they made more mistakes. Because they actually talked about them.

Her earlier research in hospitals had shown the same pattern: wards with highly competent nursing teams reported ten times as many medication errors as weaker wards. Closer analysis made it clear: it wasn't more errors. It was more conversation about errors.

From this, Edmondson developed the concept of team psychological safety — the shared belief that it is safe within the team to take an interpersonal risk. To ask a question. To admit a mistake. To voice a half-formed idea. To disagree with a superior.

Project Aristotle: When data is louder than ideology

Google didn't take this on faith. Over two years, its people analytics team studied more than 250 variables across 180 internal teams in Project Aristotle. Experience. Personality mix. Domain expertise. Who eats lunch with whom.

The result disappointed many — because it wasn't dramatic. Five factors distinguished high-performing teams from average ones. At the top: psychological safety. The other four (dependability, structure and clarity, meaning, impact) only mattered when the first was present.

Put differently: when people cannot speak up, every other factor loses its leverage.

What this looks like in money

The 2025 Gallup Engagement Index for Germany, published in March 2026, puts the economic cost of inner resignation at €119.2 to €142.3 billion per year. Only 10% of employees report high emotional engagement. 77% are doing the bare minimum — what Germans call Dienst nach Vorschrift, "service to the letter."

These numbers are not an HR statistic. They are a line item.

And they connect directly to the central question of psychological safety: do I sense it makes a difference if I bring myself in? Is it worth raising concerns? Will my contribution be heard, or will it just cost me energy?

The most common misreading

Many leaders confuse psychological safety with consensus. With harmony. With "everyone gets a turn to speak."

That is not what Edmondson is describing.

Psychological safety doesn't mean things are pleasant. It means it is possible to be honest — even when it gets uncomfortable. Friction is not the opposite of safety. It is the result. In a team with high psychological safety there is more disagreement, not less. More questions. More half-formed thinking. More mistakes named.

The system learns — because it gets information that other systems hide.

Three indicators you can actually observe

If you want to know whether your leadership team has psychological safety, you don't need an engagement survey. Three observations are enough.

First: how fast does bad news travel? Not the planned escalations. The unplanned ones. The uncomfortable ones. If risks only surface when they can no longer be hidden, the system isn't safe — it's quiet.

Second: who actually speaks in meetings? When the same three people lead the conversation while seven others nod or stay silent, that's not a mandate problem. It's a safety problem.

Third: what happens after someone disagrees? A leader's first reaction to criticism determines whether that criticism ever returns. One defensive response can take months to repair — if it's repairable at all.

What leaders can actually do

Psychological safety can't be rolled out like a new tool. It emerges in micro-moments — which is exactly what makes it hard to scale.

What works:

— Showing your own uncertainty before you expect it from others. Edmondson calls this leading with curiosity. A leader without finished answers, who says so, gives others permission to do the same.

— Training the reaction to disagreement. Not the arguments — the reaction. The breath before the reply. The question instead of the rebuttal.

— Institutionalising learning. Retrospectives, after-action reviews, pre-mortems are not method theatre. They are the infrastructure where safety gets practised.

The business case underneath

When people don't speak, the system can't learn. When the system can't learn, it loses. Not in a single quarter. But reliably, over time.

Psychological safety is therefore not a culture programme. It is the condition under which the information that drives multi-million decisions is actually valid.

This is not a soft topic. It is risk management — with different vocabulary.