Team building
Trust isn't built by an activity. It's built by people taking a small risk in front of each other and finding out that nothing bad happens. A good exercise creates that moment on purpose, in an hour, instead of waiting two years for it to happen by accident. Below are twelve, each with the debrief question that does the real work.
Most lists give you an activity name, a line claiming it builds trust, then the next one. The activity is the easy part. What decides whether anything survives Monday morning is the question you ask afterwards, whether people had a real choice about taking part, and whether the task resembles their actual work.
If you have twenty minutes, run a short activity and a long debrief. Never the reverse.
| Activity | Group size | Time | What it surfaces |
|---|---|---|---|
| Working style user manual | Any, in pairs | 20 min | Hidden preferences |
| Strength spotting round | 4 to 10 | 15 min | Who goes unnoticed |
| The leader goes first | Any | 10 min | Whether candour is safe here |
| Anonymous trust dial | 6 to 30 | 15 min | Stated versus felt trust |
| Fears in a hat | 6 to 20 | 40 min | Worries everyone holds privately |
| More of / less of pairs | Any, in pairs | 45 min | Feedback being avoided |
| Peer consulting circle | 5 to 8 | 60 to 90 min | Willingness to admit being stuck |
| Willow in the wind | 7 to 12 | 20 min | Physical reliance, carefully |
| Minefield | 8 to 20 | 40 min | Instructions under pressure |
| Blind square | 6 to 15 | 30 min | Who leads when nobody can see |
| Shared working agreement | Any, remote | 45 min | Assumptions about availability |
| Open questions with the manager | Any, remote | 30 min | What nobody asks normally |
Each person writes half a page: how I prefer to get feedback, how you can tell I'm stressed, what I'm bad at, the fastest way to get an answer from me. Ten minutes writing, ten swapping in pairs. It makes explicit what colleagues otherwise learn by getting it wrong. Keep every prompt about work, never private life.
Debrief: "What did you read that would have saved you a difficult conversation three months ago?"
Go round the group. Each person names one specific thing another named person does well, with an example. Not "she's great", but "she rewrote the client note on Thursday and stopped an escalation". Track who hasn't been named and direct the last few. Insist on specifics or you get empty compliments. The quiet, reliable ones are usually those nobody has said this about in years.
Debrief: "Which of those did you not know other people had noticed?"
The most senior person describes, in two minutes, something they got wrong recently and what it cost. Real, recent, unflattering. Then the meeting carries on as normal. That is the whole exercise. It fails when the admission is a humblebrag about working too hard. Repeat monthly rather than once.
Debrief: none on the day. Two weeks later, ask whether anything people would have hidden got raised earlier.
Anonymous poll, four statements scored one to five: I can raise a concern without it being held against me; I get the information I need; people keep their commitments; I would admit a serious mistake here. Show the spread, not the average. A team averaging 3.5 with answers from 1 to 5 doesn't have mild trust issues, it has people in different realities.
Debrief: "Someone here scored that a two. What would need to change for them to score it a four?"
Everyone writes one anonymous worry about the team, the project or the year ahead on an identical slip of paper. Collect, shuffle, redistribute. Each person reads the slip they drew aloud as if it were their own.
Debrief: "Which of these were you surprised somebody else was carrying too?"
Rotating pairs, seven minutes each way. Each names one thing they want more of from the other and one thing they want less of, with an example. The receiver may only ask clarifying questions. No defending, no trading. Keep it on behaviour, not character. Never run it where a grievance is unresolved, or where one person takes feedback from six in a row. Our guide to constructive feedback has the phrasing.
Debrief: "What did you hear more than once today?"
One person describes a real, unresolved work problem for five minutes. The group asks questions only, no advice, for ten. Then it offers options while the presenter takes notes in silence, before closing with what they will try. Twenty minutes each. Admitting you are stuck in front of colleagues is a real risk, and the group's response settles whether it was safe. Police the questions-only phase, or it becomes advice in ninety seconds.
Debrief: "What did the questions do that the advice didn't?"
A circle of seven to twelve, one person in the middle with feet together and arms crossed, leaning and being passed gently around by the group's hands. Spotters set their stance, the faller asks "ready?", the circle answers before anyone moves.
This is the honest version of the trust fall. The backward fall from a table is theatre, carries real injury risk, and someone who catches you once doesn't become someone you'd tell about a missed deadline. Make it visibly optional, and expect back, neck and pregnancy-related reasons for sitting out that nobody should have to explain.
Debrief: "Who found it harder to be caught than to catch, and does that match how you are at work?"
Scatter objects across a marked floor area. One person crosses blindfolded, guided by voice alone from outside the area. Run several pairs at once so the noise forces precision. Fifteen minutes running, twenty-five debriefing.
Vague guides say "a bit left". Good ones say "stop, quarter turn right, three small steps", then check they were understood: active listening made audible.
Debrief: "What did your guide do that made you trust them, specifically?"
The group holds a loop of rope, everyone blindfolded, and must form a perfect square, then agree they have finished. Twenty minutes maximum. Nobody can see, so leadership has to be claimed and accepted out loud, which makes the usual power dynamics visible. Watch for the good idea nobody hears, and the moment the group declares victory unchecked.
Debrief: "Who ended up leading, and how did the rest of you decide to follow them?"
Four columns on a shared board: expected response times, meetings that need cameras on, how we flag being overloaded, how we disagree in writing. Everyone adds notes silently for ten minutes, then the group clusters them and argues to one line per column. Remote trust breaks on unstated assumptions more than on personality: one person reads four hours of silence as normal, another reads it as anger. See also virtual team building.
Debrief: "Which assumption were you most surprised anyone held?"
Questions submitted anonymously in advance, answered live and unedited for thirty minutes, in the order received, awkward ones included. "I don't know" and "I can't answer that yet, here's why" are both fine. Trust is built by the hard questions answered honestly but badly. Skipping one costs more than any answer would.
Debrief: "What did you want to ask and decide not to?", collected anonymously afterwards.
Most people rush the first question and jump to lessons, which produces neat conclusions nobody believes. Staying on "what happened" longer than feels comfortable is the biggest improvement most people can make to their facilitation.
It is also why experiential learning suits trust work. You cannot instruct a team into trusting each other. You can put them into a short, real challenge, let the pattern show itself, then talk honestly about what everyone just watched happen.
Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play Gamified learning appsThe Weekly Challenge to see the method in ten minutes.
Questions
The ones that fit your time and group, and that you can debrief properly. For a short session, a strength spotting round or an anonymous trust dial. For a longer one, fears in a hat or a peer consulting circle, where people take a real risk by admitting something unresolved. The activity matters less than the conversation afterwards.
Have the most senior person go first with something genuinely unflattering, then respond well when someone else follows. Trust moves at the speed of demonstrated safety, not at the speed of activities. One well-handled admission in a normal meeting usually does more than a half-day workshop.
Not really. The backward fall from a height carries real injury risk, excludes anyone with a back, neck or balance issue, and proves only that colleagues won't let you hit the floor while being watched. That isn't the same as trusting someone with bad news. A supported lean in a circle, properly debriefed, gives you the same conversation with far less risk.
Make the unwritten rules written. Agree expected response times, when cameras are needed, how people signal overload and how disagreement happens in text. Most remote mistrust comes from misreading silence, not from a lack of social time. Then add regular, unedited question sessions with the manager.
Months of consistent behaviour, not one session. What a good activity does is compress the first step, giving people early evidence that candour is survivable here. Whether it holds depends entirely on what happens the next time someone raises something inconvenient.
Work through what happened, so what, now what. Start with observable facts before anyone interprets them, then ask what the pattern says about how the team really operates, then agree one specific change with a named owner and a date. Skipping the first question produces tidy conclusions nobody believes.
Keep reading
Get in touch
Book a 30-minute discovery call and we'll shape an experiential programme around your goals.
Book a discovery call →