The retreat that changed nothing
A company, worried about a team that never quite gets going, does what gets done: it books a retreat. A day in the country, trust exercises —the classic "let yourself fall and your teammates catch you"—, an energetic facilitator, lunch in the open air. People come back happy, with a T-shirt and a group photo. For two weeks there is better vibe in the hallways. Within a month, the team works exactly as before. The photo stays on the wall. The problem, intact.
The scene holds the central error about team development: confusing emotional bonding with performance capability. The retreat generates the first —which dissipates— and does not touch the second. The most practical question remains: can those levers be pulled on purpose? Which interventions really work —and which are expensive theater? The answer, again uncomfortable for common sense: what pays off is the structured and somewhat boring —setting goals, clarifying roles, reviewing what was done—, not the thrill of the kayak.
Teams are made, not born
The most famous model is Bruce Tuckman's (1965): forming, storming, norming and performing (later adjourning was added). It is a useful map, on one condition: take it as vocabulary, not as law. The model did not come out of an experiment but out of a review; when it was revisited in 1977, of twenty-two studies examined only one had been designed to test the sequence. Real teams advance, fall back and return to "storming" every time a member changes.
What works and what is theater
When you measure —with meta-analyses, not with retreat satisfaction surveys— which interventions really improve performance, a clear pattern appears: the structured pays off; the purely emotional does not.
Team training —systematically teaching people to coordinate, communicate and monitor one another— has robust backing: the meta-analysis by Salas and colleagues (2008), over 93 measurements in 2,650 teams, found that it explains around 20% of the variance in performance. Team building —"bringing the team together"— also has an effect, but a more modest (correlation ≈ 0.31) and uneven one: what moves performance is role clarification and goal setting; not interpersonal dynamics. An earlier meta-analysis said it without anesthesia: only interventions focused on clarifying roles reliably increased performance.
Why the retreat still "feels" good. If the interpersonal does not move performance, why does everyone come out happy? Because it measures something else: bonding raises satisfaction and climate —real, but distinct from performance—. The problem is not holding a team barbecue; it is believing that is enough. Cohesion helps, but it is the consequence of working well together, not its cause. Structure first; the bond comes behind.
The cheapest (and most ignored) intervention: the debrief
After a project, an incident or an important meeting, the team sits down for fifteen minutes to answer four questions: what did we expect, what happened, why was it different, and what do we do the same or differently next time. It is what the military call an After-Action Review. The meta-analysis by Tannenbaum and Cerasoli (2013), over 46 samples, found that structured debriefs improve performance by around 25%, with an effect size (d) of 0.67 —large—, in teams and individuals, in simulations and in real work. It is, probably, the cheapest performance improvement there is: it requires no consultants and no budget. It requires discipline and psychological safety.
The launch and the conversations no one has
The launch —the first meetings— is a moment of enormous leverage and almost always wasted. The GRPI model orders the start-up: goals first, then roles, then processes and relationships last. Most of the conflicts that look interpersonal are poorly defined goals or roles disguised as a clash of egos.
There is a preventive practice, little used and powerful, documented by Toegel and Barsoux: having, before working under pressure, conversations about the differences in how each person sees, acts, speaks, thinks and feels about work —whether a promise is a firm commitment or an aspiration, whether silence is reflection or disagreement—. Bringing those differences to light when there is no conflict "immunizes" the team. Preventing a conflict is easier than resolving it.
Sustaining: teams no longer stay still
Building a team is an event; sustaining it, a process. Haas and Mortensen describe 4-D teams: more diverse, dispersed, digital and dynamic. A team whose membership rotates cannot define its rules just once: it has to re-onboard those who arrive and repeat the norms. To Hackman's three conditions, the authors add a fourth that distance makes critical: a shared mindset, against the tendency to fracture into "us" and "them" and to operate on incomplete information. And sustaining requires measuring: a simple periodic check-up —rating the team on its conditions and on three criteria of effectiveness— catches problems while they are still cheap to fix.
How much it is worth doing well
The most forceful answer comes from Gallup, which maintains a meta-analysis on team engagement. Its 2024 edition is of a scale hard to overstate: 183,806 work units, more than 3.3 million employees, 347 organizations across 90 countries, with an engagement–performance correlation of 0.49. Comparing the top quartile with the bottom:
On the other side of the ledger is what it costs not to do it: according to a CPP study, employees in the United States spent some 2.8 hours a week dealing with conflict —the equivalent, extrapolated, of some US$359 billion in paid hours a year—. Teams working badly is not free; the bill does not arrive on a line you can see.
| Intervention | What it is | Evidence | When |
|---|---|---|---|
| Design and launch | Purpose, people, roles, norms before starting | 60-30-10 rule ‡; GRPI | When forming the team |
| Team training | Coordination, communication, mutual monitoring | ≈20% of the performance variance | New or underperforming teams |
| Role clarification | Agreeing who decides and answers for what | The component that most moves performance | Always; when facing "interpersonal" friction |
| Structured debrief | Reviewing what happened and what to adjust, 15–20 min | +25% performance (d = 0.67) | After each milestone, project or incident |
| Preventive conversations | Making style differences explicit before conflict | Prevention > resolution | At the start; with diverse teams |
| Periodic check-up | Measuring conditions and effectiveness every so often | Early detection | Continuously |
‡ a practitioner heuristic; the rest, meta-analyses. The effect sizes are not comparable with one another.
Open questions
- When you wanted to "improve the team", did you invest in structure —roles, goals, training— or in an emotional event that felt good and changed nothing?
- Does your team run structured debriefs after its milestones —or move on to the next project without distilling what it learned?
- When you built or relaunched the team, did you devote the first meetings to spelling out purpose, roles and rules?
- When facing "people" friction, do you first check whether the goals and roles are clear?
- If your team has people coming and going, do you re-onboard and repeat the norms with the newcomers?
- Do you measure, even with a simple thermometer, how your team stands on the conditions that enable it?
- Do you have an idea, even a rough one, of how much a team that does not work costs you today?
None of these interventions is expensive or exotic. All of them require the same thing: treating the team as a system that is designed and maintained, not as a group of people from whom magic is expected.
We close where we began, from the other side. We dismantled the myth (performance is designed), gave the levers names and numbers, and saw that they can indeed be pulled —that the fifteen-minute debrief pays off more than the day-long retreat, and that doing it well is repaid in profitability, safety and people who stay—. The thread through the three parts is a single one: a high-performance team is not a stroke of luck or a group photo; it is a system someone installs, measures and sustains. The aspirational, looked at closely, turns out to be engineering. And engineering can be learned.
Do you have teams you expect more from than they deliver?
32sur designs and launches teams, clarifies roles and goals, installs debrief and psychological-safety routines, and runs periodic check-ups that catch problems while they are still cheap. We don't run kayak retreats: we install the conditions the evidence associates with performance, and we stay until they work.
References
How to read the figures. An effect size (Cohen’s d) measures how large a difference is (~0.2 small, ~0.5 moderate, ~0.8 large); the debriefs’ d = 0.67 is a large effect. A correlation (ρ) runs from 0 to 1. Explaining 20% of the variance means that factor accounts for a fifth of the performance differences between teams.
- Tuckman, B. W., "Developmental Sequence in Small Groups", Psychological Bulletin, 63(6), 1965; and Tuckman, B. W. and Jensen, M. A. C., "Stages of Small-Group Development Revisited", Group & Organization Studies, 2(4), 1977.
- Salas, E., DiazGranados, D., Klein, C. et al., "Does Team Training Improve Team Performance? A Meta-Analysis", Human Factors, 50(6), 2008 — 93 measurements in 2,650 teams; ≈20% of the variance.
- Klein, C., DiazGranados, D., Salas, E. et al., "Does Team Building Work?", Small Group Research, 40(2), 2009 — moderate effect (ρ ≈ 0.31).
- Salas, E., Rozell, D., Mullen, B. and Driskell, J. E., "The Effect of Team Building on Performance", Small Group Research, 30(3), 1999 — only role clarification reliably increased performance.
- Tannenbaum, S. I. and Cerasoli, C. P., "Do Team and Individual Debriefs Enhance Performance? A Meta-Analysis", Human Factors, 55(1), 2013 — debriefs improve performance ≈25% (d = 0.67).
- Beckhard, R., "Optimizing Team Building Efforts", Journal of Contemporary Business, 1(3), 1972 — the GRPI model.
- Toegel, G. and Barsoux, J.-L., "How to Preempt Team Conflict", Harvard Business Review, June 2016 — the preventive conversations.
- Haas, M. and Mortensen, M., "The Secrets of Great Teamwork", Harvard Business Review, June 2016 — 4-D teams, the shared mindset and the periodic check-up; Hackman, J. R., Leading Teams, 2002.
- Harter, J. K. et al., Q12 Meta-Analysis: 11th Edition, Gallup, 2024 — 183,806 units across 90 countries; top vs. bottom quartile: +23% profitability, −78% absenteeism, −63% safety, −32% defects.
- CPP Inc., Workplace Conflict and How Businesses Can Harness It to Thrive, 2008 — ≈2.8 hours per week per employee in conflict; ≈US$359 billion in paid hours a year.