Team building
Most communication activities fail for the same reason: someone runs the game, everyone laughs, and nobody asks what just happened. The activity is not the training. The debrief is the training. Here are fourteen exercises we use, with the steps, the timing, and the question to ask when the fun stops.
A communication exercise does one thing: it shrinks a team's normal behaviour into fifteen minutes so it becomes visible. Whoever interrupts in the exercise interrupts in planning meetings. Whoever nods and then does something different is doing that on Thursday too. None of this is news to the team. It is just unspeakable at work, and an exercise makes it speakable.
If you are the manager, consider not running it yourself. People edit themselves in front of whoever writes their review, so a colleague from another team or an outside facilitator buys you honesty.
| Activity | Group size | Time | What it surfaces |
|---|---|---|---|
| Back-to-back drawing | Pairs, any number | 15 min | Assumed shared vocabulary |
| One word, one picture | 4-30 | 10 min | Undefined shared words |
| Yes-but, yes-and | 6-24 | 15 min | How ideas get killed early |
| Instruction relay | 6-20 | 15 min | How briefs degrade down a chain |
| The listening triad | Multiples of 3 | 30-40 min | Whether anyone truly listens |
| Paraphrase gate | 4-12 | 20-30 min | Debating misunderstood positions |
| Questions only | 4-16 | 20 min | The advice reflex |
| Blind build | Teams of 4-6 | 30 min | Instruction precision under pressure |
| Silent sort | 5-15 | 15 min | Non-verbal signals, who leads |
| Minefield | Pairs, 6-30 | 30 min | Instruction clarity under noise |
| What I heard was | 4-12 | 30 min | Gap between intent and impact |
| The one-page brief | 4-20 | 40 min | Written clarity, curse of knowledge |
| Silent shared doc | 5-40, remote | 20 min | Airtime hogged by the confident |
| Two-line update | 5-25, remote | 15 min | Rambling, buried headlines |
Pairs sit back to back. One holds a simple abstract diagram, the other a blank sheet. Four minutes to get it reproduced, words only, and in round one the drawer may not speak. Run it again with a new image, questions allowed.
It shows how much of an instruction rests on words you assume mean the same to both people. 'A box in the corner' is four different pictures. Round two is the point: results improve, and the reason is questions, not better describing. Debrief: what did you assume the other could see?
Pick a word your team uses constantly and never defines: urgent, done, aligned, quality, soon. Everyone writes a private one-sentence definition, no discussion. Then read them all out.
A team will produce five definitions of 'done', then realise three months of messy handovers came from that. Debrief: which of these words costs us most, and what would a shared definition change on Monday?
Give the group a light planning task, like organising an office move. Round one: every contribution starts with 'yes, but'. Four minutes. Round two: same task, everything starts with 'yes, and'.
Round one collapses into gloom, which is the finding. Neither round models real work, and say so: 'yes, and' with no challenge is how bad plans get funded. The question is whether your team reaches round one before anyone finishes a sentence.
One person reads a six-sentence written brief containing a deadline, a number and an exception, then passes it verbally down a chain of five or six. The last person says it aloud. Read the original.
A work-shaped brief makes the debrief land harder than the usual silly story. What survives is the narrative; what dies is the exception and the number. Ask: where does a brief pass through five people here?
The ones people resist and remember. They need a room where nobody is checking a phone.
Groups of three: speaker, listener, observer. The speaker talks for three minutes about a real but low-stakes work problem. The listener may not offer advice, opinions or their own story, only questions and reflecting back. The observer notes what the listener did, then feeds back for two minutes. Rotate.
This is the one that changes behaviour, because most people find they cannot last ninety seconds without giving advice. Debrief speakers first: what did the listener do that made you say more? See active listening.
Take a genuine open decision the team faces. Rule: before stating your own view, summarise the previous speaker's point to their satisfaction. They say yes or no. If no, try again.
Maddening for ten minutes, then something shifts. Contributions drop in number and rise in quality, because restating an argument you disagree with is hard enough that people stop performing and start listening. Debrief: how often was your paraphrase wrong first time?
Someone presents a problem for two minutes. For eight minutes the group may only ask questions. No suggestions, no 'have you tried', no statements dressed as questions. Someone calls those out.
The disguised advice is nearly impossible to stop. The value shows in what the presenter says at the end, usually 'I think I know what I'm going to do'. Instant solutions cost you the thinking of the person closest to the problem.
Two ends of a room. Team A sees a model made of bricks or blocks but cannot touch materials. Team B has identical materials but cannot see the model. They communicate by voice or one runner. Thirty minutes, then compare.
Watch for the moment the instruction-giver speeds up out of frustration and the builders stop asking questions because it feels like admitting weakness. Debrief: when did asking a question start to feel expensive, and who made it so?
The group arranges itself in a line by something knowable but unspoken: birthday month, years at the company, distance travelled to work. No talking, no writing, five minutes.
It levels the ground, because verbal fluency is removed. Debrief: who organised, and were they the people who organise when talking is allowed? The line-up assumes everyone can stand and move, so run a seated card version if not.
Scatter objects across a floor area. One person crosses with eyes closed while a partner guides them by voice from the edge. Run two pairs at once and the difficulty becomes noise, not navigation.
The lesson is not trust, whatever the internet says. It is that vague instructions fail under load and precise ones ('two small steps forward, stop') work. Debrief: what changed in your language when it stopped working? Never blindfold anyone who has not agreed, and offer an eyes-open version.
Pairs. Each gives the other one small, genuine piece of feedback about working together, phrased as a specific behaviour. The receiver's only reply is 'What I heard was...' then 'Is that what you meant?' No defence, no context. Swap.
The most demanding item here, and it needs a group with some psychological safety already. The gap between what people meant and what landed is often startling. Debrief the gap, never the content. Pair with giving constructive feedback.
Everyone writes a one-page brief for a task they own, aimed at someone who has never done it. Twenty minutes. Then swap with someone from another function, who marks every point where they would have to ask a question.
The marks cluster around jargon and around steps so obvious to the writer they went unwritten. That is the curse of knowledge, made countable. Debrief: which of our handovers would fail this test?
Video calls flatten the signals a room gives free. See also virtual team building activities.
Pose one question. Everyone writes silently in a shared document for four minutes, no talking and no reading others' answers. Two minutes reading. Then discuss, starting with points that appeared twice.
Silent writing before speaking is the most reliable fix for the call where three confident people fill fifty minutes, and it gives anyone working in a second language time to compose. Debrief: whose ideas reached the document that would never have reached the conversation?
Each person updates in exactly two sentences: what changed, and what you need from this group. Someone times it. Anything longer is stopped politely and booked separately.
Run it three weeks before judging. Week one is awkward, week two efficient, and by week three people draft their two lines in advance, which is the gain. Debrief after week one: what did you leave out, and did anyone miss it?
These are small instances of experiential learning: do something, look at what happened, decide what to change. The doing is cheap and the looking holds the value, which is why fifteen minutes with a real debrief beats two hours without one.
If the pattern is bigger than a communication habit, read Tuckman's stages next. A team that cannot listen in the triad may not have a listening problem. It may be storming, and the exercise showed you where.
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 reliable ones are back-to-back drawing, the listening triad, blind build, a paraphrase-first discussion rule, and a silent written round before any group discussion. What makes them good is not the format but the debrief: each one should end with a question that connects what happened to how the team works on a normal Tuesday.
Plan roughly half activity, half conversation. A ten-minute warm-up needs ten minutes of debrief. A thirty-minute build needs twenty to thirty. Sessions that run the activity and skip the discussion are the most common reason teams say team building does not work.
One word, one picture takes ten minutes and exposes terms the team uses without a shared definition. A silent written round before discussion takes six. The two-line update reshapes a whole status meeting. All three fit inside a normal agenda without needing a workshop.
Anything that separates thinking from speaking. Silent writing in a shared document before discussion, strict two-sentence updates, and describe-and-draw run over screen share all work well on video. Avoid formats that depend on reading a room, since a gallery of small faces will not give you that.
Three questions in order: what happened, what did you notice about how we did it, and where does that pattern show up in our real work. Ask the quieter participants and any observers first, keep your own view until last, and finish by naming one change with an owner and a date.
Compulsory participation, questions that push personal disclosure about family, money, health or beliefs, competitive formats that reward the loudest voice, and physical activities with no seated alternative. Also avoid running an exercise whose conclusion you have already written, because the group will simply perform it back to you.
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 →