# World Café Method: 90-Minute Agenda and 15 Facilitation Questions

World Café method: a 90-minute agenda, 15 questions, facilitation safeguards, and a practical framework for turning discussion into action.

- Canonical URL: https://www.harmate.com/en/blog/world-café-method-90-minute-agenda-and-15-facilitation-questions
- Author: Harmate Team
- Published: 2026-08-15
- Updated: 2026-10-18T00:44:08.800064+00:00
- Language: en

## Content

# World Café Method: 90-Minute Agenda and 15 Facilitation Questions

The World Café method helps a large group work through a series of small-table conversations. Participants discuss a question in groups of four or five, move to a new table between rounds, and finally harvest the ideas, tensions, and possible actions that emerged.

When the design is sound, the format connects experiences that rarely meet. When it is not, it produces decorated tablecloths, vague consensus, and no decision.

This guide provides a 90-minute agenda, 15 adaptable questions, and a practical harvesting framework. The goal is not participation for its own sake. It is to turn conversation into usable material.

## What is a World Café?

The model documented by the [World Café Community Foundation](https://theworldcafe.com/key-concepts-resources/world-cafe-method/) combines small tables, several conversational rounds, a table host who stays in place, and a final collective harvest.

Movement is part of the method. Each new group encounters the previous group’s notes, then extends, challenges, or connects them to a different experience. A World Café is therefore neither a collection of unrelated breakout rooms nor a disguised vote.

It works particularly well when you need to:

- explore an open question;
- connect functions, communities, or levels of seniority;
- surface tensions and common ground before a decision;
- prepare an action plan, workshop, or service change;
- capture the language participants actually use.

It is a weaker choice when a binary decision is required immediately, strict confidentiality is essential, or the output must be statistically representative. A World Café can reveal perspectives and hypotheses. It does not turn a self-selected room into a representative sample.

## Prepare the World Café: six decisions to make first

### 1. Define the required outcome

Finish this sentence: “At the end, we need to be able to…”

A useful answer names an observable output: identify three priority barriers, choose issues for further research, generate options, or define first actions. “Create collective intelligence” is an aspiration, not a deliverable.

The same principle applies to an [actionable questionnaire](/en/blog/actionable-questionnaires-start-with-the-decision-not-the-questions): begin with the decision that needs support, then design the questions.

### 2. Choose participants and table composition

The official method recommends four people per table, with five as a maximum. That leaves room for each voice while preserving enough variety for ideas to develop.

Do not automatically seat friends or colleagues together. If the purpose is cross-pollination, mix the experiences that matter. If power differences are strong, ask whether people can safely disagree with a manager. In that situation, more homogeneous tables in the first round, followed by deliberate mixing, may create a safer progression.

### 3. Build a three-question arc

Three rounds should not be three versions of the same prompt. Move the reasoning forward:

1. understand the current situation;
2. explore causes or possibilities;
3. define a practical next step.

Questions should be open, clear without a long explanation, and focused enough to prevent parallel conversations. The principles used to write [neutral open-ended questions](/en/blog/open-ended-questions-get-actionable-data-without-bias) also apply here.

### 4. Brief the table hosts

The host stays while other participants move. Their job is not to defend the previous group’s work. They welcome the new group, summarize without closing the discussion, point to the existing notes, and invite extension or disagreement.

Recent methodological research emphasizes the importance of host preparation when a World Café is also expected to produce usable data ([Klatt et al., 2025](https://doi.org/10.1017/pds.2025.10221)). Give hosts a simple task: distinguish recurring ideas, disagreements, and unanswered questions.

### 5. Define the evidence you will keep

Decide in advance whether you need exact phrases, themes, decisions, questions, photographs, or a structured summary. Without a shared template, each table records a different kind of information and the harvest becomes difficult to compare.

A simple table sheet can contain four areas:

- observations and examples;
- proposed causes or explanations;
- disagreements and reservations;
- actions or experiments to test.

### 6. Explain the decision rule

The World Café opens and structures discussion; it does not make the decision automatically. Say who will decide afterward, which criteria will be used, and when participants will see the outcome. This prevents participation from becoming theatre.

To plan that feedback loop, the [stakeholder consultation guide](/en/blog/stakeholder-consultation-deciding-without-false-consensus) distinguishes participants’ contributions, the eventual decisions, and the reasons that need to be reported back.

## Practical facilitation brief: adapt the format to your workshop

Use this copy-ready brief before inviting participants. It keeps the World Café’s conversation design connected to a decision while leaving facilitation choices to the people who know the room.

```text
WORKSHOP BRIEF — [title / date]
Purpose: [why these people are being brought together]
Decision or work product: [what the organiser will decide, draft, or investigate afterward]
In scope / out of scope: [topics and boundaries]
Participants: [roles or perspectives needed; invitation and access needs]
Format and reason: [World Café / another format; why movement between tables helps]
Questions by round: [understand] / [explore] / [act]
People and roles: [lead facilitator] / [table hosts] / [accessibility or note-taking support]
Timing and room: [arrival, rounds, transitions, harvest, breaks; room or online setup]
Participation choices: [pass, write instead of speak, take a break, leave a question unanswered]
Capture: [what is recorded, by whom, how quotations are attributed or anonymised]
Use and access: [who will read notes; how they inform a decision; retention or deletion plan]
Close the loop: [who will send the summary, by when, and how participants can correct it]
```

### A 10-minute host briefing

Before participants arrive, walk hosts through four instructions: (1) welcome each new group and give a short, neutral account of the notes already on the table; (2) invite people to add, question, or disagree, without asking anyone to disclose a personal experience; (3) record examples, interpretations, tensions, and unanswered questions separately; and (4) stop at the agreed time, leaving the next decision to the named organiser. Hosts should not diagnose people, resolve a disagreement on behalf of the group, or turn the table into a vote.

### Adapt the room without changing the purpose

For a sensitive subject, explain what will be written down and who will see it before discussion starts. Let participants contribute in writing, pass, or step out; do not photograph notes or attribute quotations unless that use was explained and agreed. If reporting back could identify someone, paraphrase or omit the detail and say that the account was withheld. A World Café is a facilitated conversation, not a promise of confidentiality. If the group cannot safely discuss the subject together, use separate conversations or another format.

For a mixed in-person and remote group, assign someone to watch the remote channel, share the same question and timing, and bring remote contributions into the harvest with their context intact. If this cannot be done reliably, run separate sessions and compare the notes cautiously rather than implying equal participation.

### One-page harvest sheet

At each table, use the same four boxes:

| What we observed or heard | Possible explanation (still a hypothesis) |
|---|---|
| [concrete example, source or context] | [interpretation and what could challenge it] |

| Tension, exception, or unanswered question | Next step for the organiser |
|---|---|
| [include minority views; do not force agreement] | [decision, check, owner by role, or follow-up question] |

At the final harvest, ask each host to share one recurring idea, one unresolved difference, and one question or action. The organiser records which proposals are accepted, deferred, or declined and why. Send participants a short account of what was heard and what happens next; do not present the conversation as a representative measurement.

## A 90-minute World Café agenda

Use this as a starting point and adapt it to the topic and room size.

### Minutes 0–10: frame the conversation

- explain the desired outcome and how contributions will be used;
- introduce the three questions and rotation pattern;
- make it clear that ideas can be challenged without attacking people;
- allow two minutes of silent thinking before anyone speaks.

### Minutes 10–30: round one — understand

Each table explores the current situation. The host asks for concrete examples instead of collecting only general opinions. End with a two-minute written summary.

### Minutes 30–35: rotate

The host stays. Everyone else joins a new table. Avoid moving intact groups together: cross-pollination is a core part of the method.

### Minutes 35–55: round two — explore

The host summarizes the existing notes in two minutes. The new group adds, qualifies, and investigates causes or options. Disagreements are recorded rather than softened.

### Minutes 55–60: rotate again

Repeat the process with another table composition. The lead facilitator checks timing and makes sure host summaries do not turn into presentations.

### Minutes 60–80: round three — act

Participants convert what they have learned into actions, experiments, or questions that require further work. Each proposal should name the problem, the first step, and the signal that would justify continuing.

### Minutes 80–90: collective harvest

Each table reports one convergence, one tension, and one action. The facilitator clusters themes without erasing minority views. Close by naming the next step: decision, report, experiment, or follow-up survey.

## 15 World Café questions to adapt

### Questions for understanding the situation

1. When does this problem become most visible?
2. What already works and should be protected?
3. Who experiences the situation differently, and why?
4. Which recent example best captures the difficulty?
5. What information are we missing?

### Questions for opening possibilities

6. If the main constraint disappeared, what would we do differently?
7. Which practice from another team or field could inspire us?
8. Which assumption are we treating as fact without testing it?
9. What might an affected person who is not in the room say?
10. Which uncomfortable option still deserves examination?

### Questions for preparing action

11. What useful change could we test on a small scale?
12. Who must be involved before that test begins?
13. What signal would show that we are moving in the right direction?
14. Which risk should be made visible now?
15. What question still needs an answer before a decision is made?

Treat these prompts as starting points. Replace generic language with the actual context without embedding the preferred answer in the question.

## Prevent the fastest or most senior voices from taking over

Small groups do not balance themselves automatically. In idea-generation tasks, interacting groups can produce less than people who first work alone. A classic meta-analysis linked the loss to factors including group size and spoken contribution ([Mullen, Johnson, and Salas, 1991](https://doi.org/10.1207/s15324834basp1201_1)). Later experiments showed that waiting for a turn can disrupt the cognitive process of generating ideas ([Nijstad, Stroebe, and Lodewijkx, 2003](https://doi.org/10.1016/S0022-1031(03)00040-4)).

A World Café does not magically remove those mechanisms. Practical safeguards include:

- begin every round with two or three minutes of individual writing;
- use an initial pass in which each person may contribute or pass;
- let participants write an idea without waiting to speak;
- ask hosts to invite quieter voices without forcing disclosure;
- keep a visible “disagreements” area on every table sheet;
- prevent hosts from translating criticism into vague consensus.

Collective-intelligence research is still debated, but an analysis combining 22 studies found that collaboration process predicted collective performance more strongly than group composition alone ([Riedl et al., 2021](https://pmc.ncbi.nlm.nih.gov/articles/PMC8166150/)). The practical conclusion is modest: conversation quality depends on the design, not only on the perceived quality of the people in the room.

## Turn the harvest into usable output

Do not rank ideas only by the number of mentions. A repeated idea may simply be familiar, socially comfortable, or inherited by several tables from the same early influence. A minority view may reveal the most consequential risk.

Organize the harvest into five categories:

1. **convergences**: themes found across conversations;
2. **tensions**: disagreements that change interpretation;
3. **minority signals**: isolated but well-supported observations;
4. **open questions**: missing information that still needs to be gathered;
5. **next steps**: decisions, owners, tests, or operational dates.

Keep a few exact phrases so that an abstract summary does not erase participants’ experience. For a larger corpus, use a structured method for [analyzing open-ended responses](/en/blog/analyzing-200-open-responses-without-bias-a-method-for-actionable-decisions) that keeps themes, evidence, and limitations separate.

A critical review notes that adoption of the World Café as a data-collection method has outpaced rigorous evaluation of whether it achieves all of its stated aims ([Clements, Sharples, and Bishop, 2024](https://doi.org/10.53841/bpsopo.2024.3.1.6)). Do not present the harvest as representative measurement. For sensitive topics or consequential decisions, triangulate it with individual interviews, a questionnaire, or observable data.

## Mistakes that undermine the method

- **A question that is too broad:** “How can we improve the company?” produces output that cannot be prioritized.
- **Three identical rounds:** participants repeat the same ideas with new neighbors.
- **An overactive host:** the host becomes the owner of the table and shapes every later group.
- **A harvest without disagreement:** the summary creates false consensus.
- **Too much reporting back:** ten sequential presentations exhaust the room and repeat the same themes.
- **No stated next step:** participants conclude that their contribution will change nothing.

## Connect the workshop to what happens before and after

A World Café is more useful when it is not an isolated event. A short pre-event questionnaire can collect expectations, identify sensitive topics, and inform table composition. Afterward, themes, decisions, and unanswered questions can feed a brief and a targeted follow-up.

If the programme also includes optional introductions between participants, the [networking planning guide](/en/blog/the-end-of-random-networking-maximizing-the-relevance-of-connections) covers consent, practical constraints, and human review; it concerns those introductions, not World Café table rotations.

Harmate can support that continuity for event organizers: questionnaires before, during, or after an event; groups when they genuinely serve the workshop; and a shareable event brief. The product does not replace human facilitation or decision-making, and it does not provide ticketing, RSVP, check-in, or event logistics. The [Event Organizers profile](/en/profiles/event-organizers) describes that boundary.

After the harvest, the [workshop reporting method](/en/blog/workshop-report-how-to-show-contributions-and-what-happens-next) helps prepare a document for participants that keeps contributions, disagreements, decisions actually made and open questions distinct.

## Final checklist

- The desired outcome fits in one observable sentence.
- Each round answers a different question.
- Tables contain four or five people.
- Hosts can summarize without defending or concluding.
- Individual writing comes before discussion.
- Disagreement has a visible place.
- Every table uses the same harvesting template.
- The decision rule and next step are explicit.
- The report states the limits of representativeness.

## Frequently asked questions

### What are the seven World Café principles?

The [World Café Foundation](https://theworldcafe.com/key-concepts-resources/design-principles/) describes them as: clarify the context, create hospitable space, explore questions that matter, encourage everyone’s contribution, connect diverse perspectives, listen together for patterns, and share collective discoveries.

### How many participants do you need?

The ideal table has four people, with five as a maximum. Total attendance depends on the room, the number of hosts, and your ability to create a useful harvest. For a very small group, a structured conversation or group interview is often simpler.

### How long should a World Café last?

Ninety minutes allows three short rounds and a harvest. For a complex topic, plan two hours and protect more time for synthesis. If you have less than one hour, reduce the number of rounds instead of rushing every conversation.

### How is a World Café different from brainstorming or a focus group?

Brainstorming primarily aims to generate ideas. A focus group usually explores the perceptions of one moderated group. A World Café emphasizes participant movement and cross-pollination across several small conversations. The formats can complement one another, but they produce different evidence. A [moving debate](/en/blog/moving-debate-shift-positions-without-measuring-opinion) makes shifts in position during discussion visible without turning them into an opinion measure. The [brainstorming guide](/en/blog/brainstorming-generate-ideas-and-choose-the-next-step) covers another idea-generation format and likewise separates generating ideas from examining them.

## The essential point

A successful World Café is not measured by the number of sticky notes or the energy in the room. It produces three things: perspectives that genuinely crossed, disagreements that remain visible, and a harvest connected to an explicit next step.

The format is simple. The rigor lies in the questions, host behavior, speaking balance, and the method used to turn conversation into a decision.