Team Treasure Hunt

Apply in a new context

Team Treasure Hunt is a team conversation about one problem everyone knows, in seven questions, and the answer is often already at the table. In groups of five to seven, everyone first tells what they already do that helps, and then who manages despite the same obstacles, inside or outside the team, and what exactly that person does differently. Only then come new ideas and actions with a name and a date. It is the Liberating Structure Discovery and Action Dialogue: we gave it a name that suits a teaching team, rewrote the seven questions in plain school language and put three prompts under each one. Made for a staff meeting or training day, and for every school that knows the expertise is in the building but does not travel.

Seven questions in groups of six · ± 85 min

Put it on the board

One screen: the problem, the group size and the time per question. Only you see it; the class sees it only when you press Start.

This exact screen goes on your board. The timer runs for every group at once and you stay in control: pause, add time or move on to the next question. The problem stays on the screen at every question, and with a free account AI helps you put it in one sentence.

What the class sees on the board

11 steps · one problem, in every group at once, ± 81 min

Team Treasure Hunt

0s in · ± 81 min to go

Conversation · step 1/11

step 1/11 · Groups and roles

Agreements from the meeting are not kept, and every week we start over.

Sit in groups of five to seven. Choose who hosts and who takes notes.

3:00

one group

Groups · 3 min

Problem · 5 min

Noticing · 14 min

What works · 8 min

What stops you · 7 min

Who manages · 10 min

New · 7 min

Who does what · 9 min

Who else · 5 min

Harvest · 10 min

Invite · 3 min

For you: Settle the roles before question 1: the host does not join in.

Group size

groups of 5–7, the whole team

Duration

± 85 min

incl. briefing and look-back

Level

teams · vocational · higher ed

Materials

paper and a pen per group

Format

parallel groups, one problem

Goal

share the team's expertise

Before question 1, agree two roles in every group. The host asks the questions and does not answer them; the scribe writes down what people do, with their name. If something is unclear, the scribe asks, and the host hands the question back to the group. Without that agreement the host joins in and nobody writes anything down.

Step-by-step instructions

11 steps

Groups and roles · 3 minutes

everyone together

Sit in groups of five to seven. Choose who hosts and who takes notes. The host asks the questions and does not answer them; the scribe writes down what people do, with their name.

The problem in one sentence · 5 minutes

in the groups

Agree together which problem you are taking on today, in one sentence everyone at the table recognises. If the problem is already on the screen, this step drops out.

Question 1 · Noticing · 14 minutes

in the groups

How do YOU notice the problem is there? Everyone tells about the last time they ran into it themselves. The longest question, for a reason: only once everyone recognises the problem in their own experience are the next six questions about one thing.

Question 2 · What already works · 8 minutes

in the groups

What do you already do yourself that helps? Everyone in turn, and so concrete that it is one action.

Question 3 · What stops you · 7 minutes

in the groups

What stops you from doing that every time? Think of a time you did not, although you could have.

Question 4 · Who already manages · 10 minutes

in the groups

Do you know someone who does manage, despite those obstacles? What does that person do differently? A colleague, someone from another team or from years ago: it is about what exactly that person does.

Question 5 · A new idea · 7 minutes

in the groups

Do you have a new idea of your own for solving this? Something that did not come up at questions 2 and 4.

Question 6 · Who does what · 9 minutes

in the groups

What needs to happen to do that, and who will take it on? One action, one name, one date, and nobody gets picked.

Question 7 · Who else · 5 minutes

in the groups

Who else do we need? Names, not roles.

Harvest with everyone · 10 minutes

everyone together

The scribes share their group's main find and who takes what on.

Who we invite · 3 minutes

everyone together

Together: who do we invite to the next conversation?

Every group answers the same question at the same time. Questions 2 and 4 carry the method: that is where what the team can already do comes out. The harvest and the invitation can go: set them to zero if What, So What, Now What follows, or if the next conversation is already planned.

Why does this work?

A team discussing a problem usually looks for the answer outside itself: a training, a new rule, a colleague who should take it on. Team Treasure Hunt turns that around. Its starting point comes from the work on positive deviance: in almost every group there is someone who, with the same means and the same obstacles, still gets a better result. Three things help the conversation find that person.

What already works, first

Question 2 does not ask what is wrong, but what you already do that helps. Everyone gets a turn, including whoever says they do nothing special, and that is often where the find is: a sentence at the door, a fixed place for bags, a first task already on the board. The scribe writes down what people do, not what they think, so the sheet at the end holds actions someone else can take over.

The answer is at the table

Question 4 is the heart: do you know someone who does manage, despite the same obstacles, and what does that person do differently? That may be a colleague, someone from another team or from years ago. The prompts under it move from a name to an action: what exactly did you see that person do, and what in that is different from what you do? A team holds far more expertise than a meeting ever lets you hear; this question makes it heard.

Roles that carry the conversation

The host asks and does not answer. It looks small, but it keeps the conversation in the group: a host who also answers becomes the group's expert. The scribe makes the difference between a good conversation and one that leads somewhere, because actions with a name and a date are not forgotten. And because every group answers the same question at once, the whole team talks with everyone in an hour, not just with the loudest.

When it fits, when it doesn't

When it fits

With a problem everyone in the team knows. Students who arrive without homework, a restless last period, parents who respond only when it is too late. The more shared the problem, the more experience at the table.
When the expertise is there but does not travel. In a team where everyone solves it in their own classroom, nobody knows what the colleague next door does. This conversation is made to make that visible.
At a staff meeting or training day with time. Eighty-five minutes is half a meeting. Plan it as the meeting's topic, not as the last item on a full agenda.
When you want actions that actually happen. Question 6 ends with one action, one name and one date, and nobody gets picked. An action someone chose themselves is one that gets done.

When it doesn't

For one colleague's case. If it is about one person's situation, pick a peer-consultation method with a presenter, such as trio advice or the incident method. This conversation is about everyone's problem.
When the team cannot name the problem yet. Then run What, So What, Now What at an earlier meeting first. The answer to Now what? is the problem for this conversation.
When someone is looking for a culprit. A problem that sounds like blame ("colleagues don't check homework") makes questions 2 and 4 unsafe. Rephrase it as what everyone meets, or wait until the tension has gone.

Examples by context

One team, 6 November: First a warm-up in pairs on separate questions, a new partner for every question, found by raising your hand if you had nobody yet, as in Movement Quiz. Then the seven questions in groups, sixty minutes. After that, without a harvest, straight into What, So What, Now What.

Staff meeting: Students arrive without their homework done, and the lesson stalls. Question 4 found a colleague who always opens the first five minutes with the homework.

Department: Test results in year 9 are falling behind. Question 2 showed that two colleagues already used practice tests, without the others knowing.

Form tutors: Students who go quiet in class. The conversation ended with an agreement on one fixed question in every tutor talk.

Training day: How do we keep the last period calm? Five groups at once and the same question on the screen.

Variations

With a problem already chosen if the host brings a problem the team has talked about before, question 1 is less about searching and more about checking: does everyone recognise it in their own lessons?

With What, So What, Now What after set the harvest to zero and move straight on to What, So What, Now What after the seven questions. It harvests every group's finds better than a round of scribes.

With 1-2-4-All before Snowball (1-2-4-All) also gathers the problem from the whole team, but together with this conversation it runs well over an hour and a half. Only use it if there are two meetings.

With a warm-up in pairs separate questions in pairs, a new partner for each, found with a raised hand, as in Movement Quiz. Everyone is already talking before the groups start.

Shorter give questions 3, 5 and 7 less time. Never cut questions 1, 2 and 4: that is where the method is.

Common mistakes

The problem sounds like blame. Then question 2 becomes a defence and question 4 a ranking. Describe what you see, not who gets it wrong; in the preparation AI helps with phrasings.

The host joins in. A host who also answers becomes the group's expert, and the scribe writes down their opinion. Settle the roles before question 1.

Question 1 gets skipped. Everyone thinks they know the problem, so the conversation jumps to solutions. But without a shared picture the next six questions are about six different problems.

Question 3 becomes a round of complaints. It is about what stops YOU doing what you named at question 2, not about the organisation in general. Ask back: and what does that mean in your lesson?

Question 4 stops at a name. "Sarah is good at that" helps nobody. Keep asking what exactly Sarah does: so concrete that someone else could do it tomorrow.

Actions without a name. "We'll pay more attention to it" never happens. One action, one name, one date, and whoever volunteers.

Frequently asked questions

What exactly is Team Treasure Hunt?

A team conversation about one shared problem, in groups of five to seven that answer the same seven questions at once: how do you notice the problem, what do you already do that helps, what stops you, who manages despite everything and what do they do differently, what new idea do you have, who does what, and who else do we need. Each group has a host who asks the questions and a scribe who writes down actions and names. It is the Liberating Structure Discovery and Action Dialogue, with a name and questions for teams, in a school or beyond.

How long does Team Treasure Hunt take?

With the default settings the whole structure takes ± 85 minutes, briefing and look-back included.

The seven questions take sixty minutes by default: 14, 8, 7, 10, 7, 9 and 5. With forming the groups, agreeing the problem, the harvest and the invitation, the conversation comes to about eighty-five minutes. Whoever prepares it sets the time per question; the player counts down in every group at once. If it has to be shorter, give questions 3, 5 and 7 less time.

Why is question 1 the longest?

Because everything after it builds on it. A team that says "homework" often means five different problems. Only once everyone has named a concrete example from this week and recognises the same thing are the questions about what works and who manages about one thing. You never win that time back later.

How does AI help with the problem?

In the preparation you describe in your own words what the team keeps running into. With a free account, AI then suggests four ways to say it in one sentence, as something everyone meets and without a culprit. You pick one or write your own; it stays on the screen at every question.

Where does the method come from?

Discovery and Action Dialogue is one of the Liberating Structures by Henri Lipmanowicz and Keith McCandless. It builds on Jerry and Monique Sternin's work on positive deviance: the answer to a shared problem is often already with someone in the group who, with the same obstacles, does it differently. Liberating Structures are freely available under a Creative Commons licence; the name Team Treasure Hunt, the questions in school language and the timings on this page are LudensFlow's.

Which team structure do you pick?

Structure	Goal	Group	Duration	When it fits
Team Treasure Hunt	bring out the team's expertise on one problem	groups of 5–7	± 85 min	a problem everyone knows
What, So What, Now What	make sense of a shared experience and pick one step	groups of 4–6	± 30 min	after a training day or project
Snowball (1-2-4-All)	gather everyone's ideas, from alone to the whole group	8–32	± 18 min	the team is still looking for the problem
Trio advice	quick advice from two colleagues, in parallel trios	3 per group	± 35 min	everyone has their own question
Incident Method	analyse one incident blind, to alternatives	5–8	± 40 min	one colleague's case

Sources

Liberating Structures: Discovery and Action Dialogue (Creative Commons BY-SA): the method itself, the seven questions and their order.
Henri Lipmanowicz and Keith McCandless, The Surprising Power of Liberating Structures (Liberating Structures Press, 2013).
Richard Pascale, Jerry Sternin and Monique Sternin, The Power of Positive Deviance (Harvard Business Press, 2010): why the answer is often already in the group.

Compiled by Oksana Oliinyk · Last updated 6 November 2026

Was this page useful?

After your answer we ask three short questions. Half a minute, and it really helps us.