Google studied 180 teams and found that psychological safety was the single strongest predictor of team effectiveness, correlated with 43% of the variance in performance (Google re:Work, Project Aristotle). That finding is nearly a decade old now. Every leadership program cites it. Every People team has a slide about it.
And yet: 63% of workers still do not feel safe sharing their opinions at work (Niagara Institute, 2025). Only 26% of leaders actively foster psychological safety on their teams. The concept won. The execution lost.
Why? Because most organizations treat psychological safety like a milestone. You run the offsite. You do the trust falls (or the modern equivalent: a vulnerability circle). You check the box. Then you move on to the next quarter's priorities and wonder why the team is back to silence in standups by month three.
New research explains exactly what is going wrong.
Psychological Safety Is Perishable. The Science Now Proves It.
A 2025 study in the Journal of Business and Psychology tracked teams across multiple contexts and found something that should reshape how every leader thinks about safety: it does not accumulate passively. Time together is not enough. Psychological safety emerged from, and was sustained by, four distinct team processes: connecting, clarifying, supporting, and performing (Springer, 2025).
When teams actively engaged in all four processes, safety grew. When any process dropped off, safety decayed. The researchers described it as a "perishable resource" that can diminish just as easily as it grows.
This aligns with what Harvard's Amy Edmondson has been finding in longitudinal data: new employees start jobs with higher psychological safety than their tenured colleagues, then lose it over time. The recovery period? The data suggest it could span 20 years or more (Harvard Business School, 2024).
Read that again. Twenty years to rebuild what took months to destroy.
The Four Processes (and Why Most Teams Only Do One)
Let's break down what the research actually identified.
Connecting is the process of building interpersonal familiarity, trust, and belonging. This is the one most organizations focus on. Happy hours. Icebreakers. Donut pairings. It is real, it matters, and it is also the easiest process to mistake for the whole thing.
Clarifying means establishing shared understanding of goals, roles, and expectations. When people do not know what success looks like or who owns what, ambiguity creates anxiety. Anxiety kills candor. Teams with high clarity can disagree about how without fighting about what.
Supporting involves actively helping teammates, sharing workload, and responding constructively to mistakes. This is where the rubber meets the road. A 2024 APA survey of over 2,000 U.S. workers found that 84% rank a psychologically safe workplace among their top priorities, but only 50% say their manager actually creates it (APA, 2024). The gap between valuing support and delivering support is enormous.
Performing is the process of executing together under real conditions, learning from outcomes, and adapting. This is the process almost every team skips in their psychological safety efforts. You cannot practice performing in a workshop. You need live reps with actual pressure.
Most organizations invest heavily in connecting, occasionally in clarifying, rarely in supporting, and almost never in performing. Then they wonder why safety evaporates after the offsite glow fades.
What Decay Looks Like in Practice
Psychological safety does not collapse overnight. It leaks. Here is the typical pattern:
Month 1 post-initiative: The team feels energized. People speak up more in meetings. Someone shares a concern they had been sitting on for weeks. Progress.
Month 3: The novelty wears off. A new project creates deadline pressure. A manager, under stress, dismisses an idea without realizing it. Nobody says anything about it. The person who spoke up earlier takes a mental note.
Month 6: The team is back to pre-initiative behavior, but now with an added layer of cynicism. "We did that psychological safety thing. It didn't stick." Meanwhile, the company is already planning the next offsite.
This pattern is predictable because it follows the research: connecting without clarifying, supporting, and performing is a sugar rush. The energy is real but it has no structural foundation.
A meta-analysis published in Frontiers in Psychology found that psychological safety's impact on performance is fully mediated by learning behavior and team efficacy (Frontiers, 2020). Translation: feeling safe does not improve performance directly. Feeling safe and then actually using that safety to learn and adapt together is what moves the needle. If you build the feeling but never exercise the behavior, you get nothing.
Remote Teams Have It Worse
The decay problem accelerates in remote and hybrid environments. Without hallway conversations and shared physical space, connecting degrades first. Nearly half of female business leaders report difficulty speaking up in virtual meetings, and 1 in 5 employees feel overlooked during video calls (APA, 2024).
Remote teams need more deliberate practice, not less. Yet the standard remote playbook is fewer touchpoints, more async, and the assumption that Slack messages substitute for the real thing. They do not. Remote team trust issues compound when the four processes are not actively maintained.
The retention math is brutal: when psychological safety is high, only 3% of employees plan to quit. When it is low, that number jumps to 12% (Niagara Institute, 2025). For a 200-person company, that is the difference between losing 6 people and losing 24.
What Weekly Practice Actually Looks Like
If psychological safety is perishable, the solution is not a bigger annual investment. It is a smaller, more frequent one. You need a system that exercises all four processes on a regular cadence.
This is the design principle behind QuestWorks, the flight simulator for team dynamics. It runs teams through scenario-based challenges on its own cinematic, voice-controlled platform. Each quest forces the team to connect (collaborate in real time), clarify (align on goals mid-scenario), support (cover for each other under pressure), and perform (execute together and see what happens).
QuestDash surfaces the behavioral data: who stepped up, where communication broke down, which patterns are shifting week over week. Leaders get aggregate trends and strengths-based highlights through a weekly team health report. HeroGPT provides private AI coaching that never shares upstream. Everything is voluntary and not tied to performance reviews.
The point is not that QuestWorks is the only way to exercise these four processes. The point is that something has to exercise them, and it has to happen more than once a quarter. A comparison of simulators vs. traditional events makes the structural difference clear: events are episodes, practice is a system.
At $20/user/month with a 14-day free trial, the cost of weekly practice is lower than a single quarterly offsite. Companies spend an average of $212 per employee annually on team-building activities (TeamBonders, 2025). Most of that budget goes to one-time events that produce connecting without clarifying, supporting, or performing.
The Measurement Problem
You cannot improve what you cannot see. Most teams have no way to know whether psychological safety is growing, stable, or eroding between their annual survey cycles.
Traditional measurement tools each capture a piece: pulse surveys catch sentiment, 360s catch perceptions, personality assessments catch traits. None of them capture real-time behavioral patterns under pressure, which is where safety either holds or breaks.
New managers are especially vulnerable here. They inherit a team, assume the existing safety level is stable, and do not realize it is already in decline until someone quits or a project implodes.
Stop Treating Safety Like a Milestone
The research is clear. Psychological safety is not something you build once. It is something you maintain through active, repeated practice across four specific processes. Skip any one and the whole structure weakens.
The organizations that get this right will stop asking "have we built psychological safety?" and start asking "are we practicing it this week?" That shift, from achievement to maintenance, from milestone to habit, is the difference between teams that sustain high performance and teams that peak for a quarter and then quietly regress to the mean.
Your team felt safe six months ago. Do they feel safe now? If you do not have a system that answers that question with behavioral data, you are guessing. And the research says your guess is probably wrong.
Try it free
