Six Thinking Hats
Six Thinking Hats® is a group discussion method devised by Edward de Bono. Each coloured hat stands for a mode of thinking: facts for white, emotions for red, risks for black, benefits for yellow, new ideas for green, the management of the thinking for blue. The whole group wears the same hat at the same moment, then moves on to the next one together. The group examines a proposal from one angle at a time instead of debating it in two camps. The blue hat opens the session with the purpose and the sequence of hats. It closes the session with a synthesis.
Goal
Six Thinking Hats splits a group discussion into steps in which all participants think in the same mode. The de Bono Group describes the method as a parallel thinking process: the group separates thinking into six functions and moves from one to the next by changing hats.
The method replaces the debate in an ordinary meeting, where one participant defends a proposal and another attacks it. Facts, fears and ideas arrive mixed together. With the hats, the participant who defended the proposal also looks for its risks under the black hat. The one who attacked it also looks for its benefits under the yellow hat.
The technique prepares a decision on a proposal, a problem or a review. The group decides in the session or passes a recommendation to the body that decides. The deliverable is the hat-by-hat minutes: the contributions filed under the hat that produced them, then the synthesis of the final blue hat with its outcomes and their owners.
The PMBOK Guide cites the six hats as an example of a problem-solving method among the tools of the Lead the Team process (§2.6.2.4). Its Section 5 defines them as a technique for group discussion and creative conversation. It gives the function of each hat. It does not describe how a session runs.
Usage
When to use
- A proposal that splits the team over the analysis: both camps examine the risks together, then the benefits together.
- A decision to prepare for a committee: the facts, the risks and the options arrive separately.
- A line manager in the room: the black hat turns criticism into a contribution asked of everyone.
- Unspoken fears that hold the group back: the red hat gives them a place on the agenda.
- A review after an incident or a phase: the group separates the facts from its feelings before it draws lessons learned from them.
- An inexperienced facilitator: the announced sequence and the five rules structure the discussion on the facilitator's behalf.
- A distributed team: the facilitator shows the current hat on screen and everyone sees the same marker.
When not to use
- The group must produce many options: the green hat is only one step. Open up the options with brainstorming.
- Several options to rank against weighted criteria: the hats examine without scoring. Use a decision matrix.
- A disagreement over interests rather than over an analysis: a shared mode of thinking does not divide a budget. Move to negotiation.
Description
Parallel thinking
In parallel thinking, the whole group wears the same hat at the same moment. Zhang et al. write that by mutual agreement the group wears the hats one at a time.
A hat stands for a mode of thinking. Nobody is the team's "black hat". This rule distinguishes the technique from role play.
The blue hat has two uses. The whole group wears it at the opening and at the close like the other hats. The facilitator also performs its function for the whole session: they announce the next hat and enforce the rules. The PMBOK Guide says that blue usually facilitates the meeting. The de Bono Group makes it the control mechanism that watches over the rules. This steering function is the only one that one person performs alone.
The six hats
The PMBOK Guide and The de Bono Group describe the function of each hat. Zhang et al. propose questions for the facilitator and sequences for a debriefing. They present these sequences as examples.
| Hat | Mode of thinking | Facilitator's questions | Example sequence (Zhang et al.) |
|---|---|---|---|
| White | Facts: the information known or needed | What facts do we have? Which ones are we missing? How do we get them? | After blue, to establish the facts |
| Red | Emotions, intuitions, fears, likes and dislikes | How do you feel? How do you react to this? | After black, to find out how the group feels about it |
| Black | Caution: the risks, the difficulties, the reasons for failure | Will this work? What are its flaws? Is it accurate? | After white, to bring out the obstacles |
| Yellow | Benefits: the value of the proposal and what will make it work | What are the benefits? How does this help us? Why will this work? | After green, to support an idea |
| Green | Creativity: possibilities, alternatives, new ideas | What ideas do you have? How else could we solve this problem? | After black or blue, to look for ideas |
| Blue | Managing the thinking: the purpose, the agenda, the sequence, the outcomes, the synthesis | How do you want to approach the question? How do we sum up what we have established? | At the start and at the end of the session |
The PMBOK Guide gives black a negative point of view. The de Bono Group makes it the hat of risk management. Zhang et al. give it caution and judgement: under it the group looks for the weaknesses of a proposal in order to remedy them. Yellow counterbalances it: under yellow the group looks for the value with the same rigour.
The rules of participation
Zhang et al. report the method's rules of participation. All participants accept them before the first step.
- Each member accepts everyone's ideas.
- The group wears only one hat at a time.
- Each contribution matches the mode of thinking of the hat being worn.
- A contribution that belongs to another hat waits for that hat's turn.
- Each contribution leads to a useful outcome.
Running a session
Open under the blue hat. The group sets the purpose of the session, the agenda and the expected outcomes: a recommendation, a list of risks, a decision. The facilitator announces the rules and the sequence of hats before the first step. Zhang et al. built one sequence per type of debriefing: the quick review, the brief feedback, the emotionally charged case and the full review. The sequence therefore follows the purpose of the session. To evaluate a proposal, the facilitator places white right after the opening. Black and yellow then work on established facts.
Move the group from one hat to the next. The facilitator announces the hat and shows it. In the study by Zhang et al., the facilitator wore coloured caps and changed them at each step. All participants contribute under the announced hat.
Record each contribution under its hat. The board has one area per hat and a "parking lot" area. In the parking lot, the facilitator notes an off-hat remark under the name of the hat it belongs to. The facilitator reads it out again when that hat comes up (rule 4). The participant does not have to remember it.
Close under the blue hat. The group sums up what it has established and formulates the outcomes announced at the opening. It gives each outcome an owner and a deadline. A session without this last hat leaves one list per hat and no conclusion.
Pitfalls
The off-hat remark is the pitfall that Zhang et al. observed. In their first debriefing, the participants forgot the function of the hat the facilitator was wearing and made off-hat remarks. At the second debriefing, each cap carried the word for its mode of thinking. The participants stayed in the requested mode better. They reminded a colleague of the rule themselves. So the facilitator writes the word for the mode of thinking on each marker: "Facts", "Risks", "Benefits".
A hat assigned to a person breaks rule 2 and restores the debate that the method removes. When the facilitator gives black to the financial controller and yellow to the project manager, each one returns to their starting position under a new name. Apart from the steering function that the facilitator performs, the whole group wears every hat.
A black hat that overruns leaves the group with a list of reasons to do nothing. The de Bono Group flags this risk: black becomes a problem when the group overuses it. The facilitator places a green hat after black. Zhang et al. give this sequence to generate ideas.
A red hat spent justifying feelings loses its purpose. A participant who justifies a fear turns it into a black-hat argument. The facilitator records the fear as it is spoken. The argument waits for the black hat (rule 4).
Related techniques
| Technique | What it does | When to prefer it to the six hats |
|---|---|---|
| Brainstorming | Produces as many ideas as possible on an open question. In a Six Thinking Hats session, the green hat takes up a single short step. | The group must first open up the options. |
| Creative thinking | Brings together the techniques that shift the starting point: analogy, assumption reversal. | The green hat produces only expected variants. |
| SCAMPER | Transforms an existing object through seven questions. The six hats change the group's stance and leave the object intact. | The group is looking for variants of a product or a process. |
| Pro versus con analysis | Weighs an option in two columns. Black and yellow supply the material. | A single person must settle a simple decision. |
| Risk analysis and management | Rates each risk by its probability and its impact, then tracks it. | The risks named under the black hat must go into the project's risk register. |
AI considerations
A language model prepares each hat's questions for the topic of the session. The facilitator starts from the generic questions for each hat and obtains wordings specific to the proposal. The facilitator reviews them and removes those that steer the answer.
After the session, the model sorts raw notes under the six hats. It flags the contributions that seem off-hat. The facilitator checks each classification, because the model can confuse a red-hat fear with a black-hat risk.
Under the white hat, the model proposes a list of facts to gather. It does not supply these facts: an unsourced figure from the model looks like a fact without being one. Under the black hat, it names generic risks. It does not know the local constraints: the internal regulations, the collective labour agreement (CLA), the systems in place.
The red hat cannot be delegated. A model feels no fear about the project and does not know the team's fears.
Session notes contain names and opinions about colleagues. These opinions are personal data within the meaning of the Federal Act on Data Protection (FADP). The facilitator removes the names before submitting the notes to a public model. A provider that processes real notes on behalf of the organisation is a processor within the meaning of Art. 9 FADP: the organisation assigns this processing to it by a contract that complies with that article.
Examples
The final blue hat draws the recommendation from the minutes.
| Step | Hat | Question asked | What the group recorded |
|---|---|---|---|
| 1 | Blue | What must the session produce? | A recommendation for the 10 October steering committee meeting. Sequence: white, black, red, green, yellow, blue. |
| 2 | White | What facts do we have? Which are missing? | Customer service receives 2'400 calls a day in November. The average over the year is 900. User acceptance testing (UAT) covers 85% of the cases. Missing: the hosting provider's peak capacity. |
| 3 | Black | What can fail? | A login outage would hit the busiest month of the year. The 15% of untested cases concern the deductible change. |
| 4 | Red | How do you feel? | Customer service fears absorbing the calls of the first week alone. The project manager is confident. Two developers are weary of the postponements. |
| 5 | Green | What other options are there? | Open on 3 November to 5% of policyholders drawn at random. Open on 3 November without the deductible change. Open on 12 January. |
| 6 | Yellow | What does each option bring? | The 5% measure the actual load and limit the effect of an outage. Opening without the deductible change keeps the date announced to policyholders. 12 January avoids the peak. |
| 7 | Blue | What do we conclude? | Recommendation: open on 3 November to 5% of policyholders without the deductible change. Open to all if the error rate of portal transactions stays below 1% for one week. Condition: the hosting provider's capacity known on 8 October (project manager). Two extra staff in customer service in the first week (head of customer service). Open the deductible change when user acceptance testing covers the remaining 15% of cases (project manager). |
| Parking lot | Off-hat | Noted at step 2 and taken up at step 3 | "It's too risky in November." The facilitator read it out under black, where it became the outage risk. |
The recommendation answers the red-hat fear with extra staff in customer service.
Visualisations
The minutes, the hats and the related techniques fit into rows and columns: they are tables. A table does not show how a session unfolds. The blue hat holds both ends of the session. Between the two, the whole group wears the same hat at each step in a sequence chosen at the opening. An off-hat remark leaves the timeline and comes back to it further on. These links call for a diagram.
Cost
| Phase | Level | Justification |
|---|---|---|
| Preparation | Low | Formulate the proposal, choose the sequence, prepare six visual markers. On first use, train the facilitator: Zhang et al. trained theirs in two hours. |
| Execution | Low | One meeting of the group concerned. The session replaces the discussion that the group would have held anyway. |
| Documentation | Low | One page of minutes: one line per hat, then the blue hat's outcomes with their owners. |
Tooling
Six visual markers are enough: hats, caps or cards placed on the table. Each marker carries the name of the colour and the word for the mode of thinking. The word stays readable for a participant who has trouble telling red from green. A whiteboard divided into six areas collects the contributions. A seventh area is the parking lot for off-hat remarks. Sticky notes in six colours keep the hat visible after the notes are regrouped.
For a distributed team, one slide per hat keeps the marker on screen. Digital whiteboards such as Miro, Mural, Conceptboard or Microsoft Whiteboard reproduce the six areas and let everyone write at the same time. A word processor is enough for the minutes.
Sources
- Project Management Institute, A Guide to the Project Management Body of Knowledge (PMBOK Guide), 8th ed., PMI: the place of the six hats among the problem-solving methods of the Lead the Team process (§2.6.2.4); the definition of the technique and the function of each hat (Section 5).
- The de Bono Group, "Six Thinking Hats": parallel thinking, the function of each hat and the risk of overusing the black hat.
- Zhang, X. C., Lee, H., Rodriguez, C., Rudner, J., Papanagnou, D., "A Novel Approach to Debriefing Medical Simulations: The Six Thinking Hats", Cureus 10(4), 2018, e2543: the questions per hat, the example sequences, the sequences per type of debriefing, the rules of participation, the opening and closing under the blue hat, the word written on the visual marker. The authors present their work as a feasibility study on two medical simulation debriefings.