Strategy Workshop Facilitation: Who Speaks First Decides

Juror 4 changed his vote in about a second

In one of Reid Hastie's mock-jury studies, the volunteers wrote down their verdicts privately before deliberating. Then the jury took a straw vote around the table. Jurors 1, 2 and 3 said second-degree murder. Juror 4 had privately voted not guilty and marked the highest level of confidence on his ballot. He paused, and said “Second degree.” When an undecided juror asked why, juror 4 answered, “Oh, it's just obviously second degree” ¹.

Cass Sunstein and Hastie, who tell the story, add a line every strategist should tape above the whiteboard: “We have no doubt that similar scenarios play out every day in jury rooms, boardrooms, and conference rooms all over the world” ¹.

Now picture the strategy offsite. The CEO opens with her view of the market. The head of the largest unit agrees. By ten o'clock the room has a consensus, and the regional manager who knows the competitor just cut prices in two of your best territories has decided it is not the moment.

Strategy workshop facilitation is the work of designing and running a session so that a group pools what its members actually know and reaches the outcome it was convened for: a set of options, a decision, or a plan. The thesis of this page is that most of that work is about order. Get every view on paper before the most powerful person speaks, then converge by a rule everyone agreed before the day began. Get the order wrong and no amount of good content will save the session.

Write before anyone talks. The boss speaks last.

Decide what kind of workshop it is before you design it

Most executive meetings are not built to decide anything. In a survey of top management teams at 187 companies with market capitalizations of at least $1 billion, Michael Mankins found that more than 65% of leadership meetings were held for “information sharing,” “group input,” or “group discussion,” and only 12% of executives believed their meetings consistently produced decisions on important strategic or organizational issues ². The strategy sessions, he wrote, were most commonly off-site brainstorming events that produce few tangible outputs ².

A facilitator's first question is therefore not about exercises. It is: what will exist at the end of the day that does not exist now? Three answers cover most strategy workshops.

Do not try to do the first two in one sitting if the stakes are high. Textron organized its strategic-issue reviews around “facts, alternatives, and choices” in two half-day sessions per issue: the first to agree the facts and a limited set of viable alternatives, the second to evaluate them and choose ³. Mankins reports that ABN AMRO, Cadbury Schweppes and Boeing likewise often held a meeting to discuss alternatives before the one that approved a course of action ². The workshop design canvas turns this choice into the first box you fill in.

Four ways a room goes wrong

Sunstein and Hastie trace group failure to two forces. Informational signals: people reasonably learn from what others say, so a wrong signal from some members misleads the rest. Reputational pressure: people go quiet, or shift their view, to avoid disapproval, which costs more when the person disapproving has power ¹. Those forces produce four problems ¹:

Two findings matter most for strategy work. Polarization happens on matters of fact, not only values. Asked on a scale of zero to eight how likely a product is to sell a certain number of units in Europe, a group whose median answer before discussion is five will tend to go up; one whose median is three will tend to go down ¹. And the planning fallacy, the habit of underestimating time and cost, is aggravated in groups ¹. A workshop that ends with a more confident forecast than anyone walked in with has not necessarily learned anything.

The facilitator's moves, in order

Every move below does one of two jobs: it keeps people's views independent until they are visible, or it makes the convergence explicit. Use them in roughly this order.

1. Open with the question and the ruleState the question the room is answering, what the session will produce, who decides if the room does not agree, and how. Write all four where everyone can see them. If the decider is in the room, they confirm it out loud.
2. Write before talkingBefore any discussion of an option, a forecast or a risk, everyone writes their own view in silence. In Gary Klein's premortem, people independently write down every reason they can think of for the project's failure, then the leader asks each person in turn to give one reason until all are recorded ⁴. The same pattern works for options, assumptions and estimates.
3. Widen in small groupsFor idea generation, Liberating Structures' 1-2-4-All runs about 15 minutes: one minute alone, two in pairs, five in groups of four, then about seven minutes in which a few people share with the whole room what everyone should hear ⁵. Its authors describe it as a way to surface ideas people would otherwise keep to themselves and to diminish power differentials ⁵.
4. Silence the leaderLeaders promote self-censorship by stating their views early, Sunstein and Hastie write; they can instead decline to take a firm position at the outset and ask for information only others hold ¹. Agree this with the CEO the day before, not in the room.
5. Assign roles for what people knowTell each participant, before discussion, what distinctive information they are there to bring: the customer view, the cost data, the regulatory risk. Experiments found the bias toward shared information falls when people are openly assigned specific roles ¹.
6. Attack the leading option on purposeA red team asked to build the strongest case against a plan works better than a single devil's advocate, whom the group knows is only playing a part ¹.
7. Vote privately, then talkBefore the group discusses its preference, collect individual votes or estimates in writing. A simpler variant Sunstein and Hastie describe is to deliberate first and then give final votes anonymously, which insulates people from reputational pressure ¹.
8. Close in writingRead back what was decided, who owns it, the resources it needs and when it will be reviewed. Mankins notes that the hardest thing for a top team is often agreeing on what it agreed to, and that undocumented decisions fall to the silent veto: people who disagree simply relay nothing to their teams ².

Running the convergence

Divergence is the pleasant half of a workshop. Convergence is where facilitators earn their fee, and where most of them flinch.

When a group stalls, the disagreement is almost always about one of four things, and each needs a different move. Name which one it is out loud.

• The facts. People are arguing from different numbers. Stop the debate, write the disputed fact on the wall and either settle it now or make it the first item of the next session. Textron's first session existed for exactly this ³.
• The options. One option is really three, or two options are the same plan at different sizes. Rewrite them until they differ in kind. Mankins argues for at least three real alternatives, not minor variations on one theme ².
• The criteria. People agree on the facts but weigh them differently. Make the trade-off explicit: if speed and cost conflict, which wins? Agree the criteria before scoring anything; the decision-making framework page covers the sequence.
• Who decides. The argument is really about authority. No exercise fixes this. Send it to the person named in the opening, or end the session and fix the decision rights first.

The rule for converging should be set before the day, not discovered at 4:30. Common choices are a named decider who hears the room and decides, a vote with the decider able to overrule and explain why, or consent (proceed unless someone has an objection they can state as a risk). Any of them works if it was agreed in advance. Consensus on a hard choice, used as a default, tends to produce the sentence everyone can sign and nobody can act on.

A workshop, run twice

Notice what did not change. Same people, same facts, same day. The second session had no clever exercise. It had a different order.

Inside or outside facilitator

The strategy team usually ends up facilitating its own leadership team's workshops. That is convenient and often a mistake, because the strategy team frequently has a stake in the answer: it wrote the options, or it owns the recommendation.

Daniel Kahneman, Dan Lovallo and Olivier Sibony make a related point about reviewing recommendations. Many companies, they write, expect a functional leader such as a CFO or chief strategy officer to play the challenger, but “an insider whose primary job is to critique others loses political capital quickly”; their advice is to rotate the role ⁶. The facilitator's chair has the same problem. Whoever holds it every time slowly becomes either the CEO's instrument or the person everyone works around.

Our rule of thumb, which is judgment rather than research: the strategy team can facilitate when it has no recommendation on the table. When it has one, it should present, and someone without a stake should run the room: an executive from an unaffected function, a peer from another business unit, or an outside facilitator. The stakeholder map for the decision usually shows who is genuinely neutral.

There is a second-order problem at the top. Facilitation techniques work on everyone in the room except the person who convened it. A CEO who tells the facilitator to run a fair process and then signals a preference over coffee has undone the morning. That conversation happens the day before, privately, and it covers three things: when the CEO will speak, how the decision will be made, and what the CEO will do if the room leans the other way.

The case against all this process

The strongest objection is that facilitation is theater. Senior executives are paid to decide. Silent writing, sticky notes and private votes treat them like a workshop of new hires, and a decisive CEO should simply hear the arguments and choose.

Half of that is right. Nothing on this page replaces a decider, and a facilitated session that dissolves accountability into “what the group thinks” is worse than a CEO deciding alone. That is why the opening move names who decides.

The other half misreads what the techniques are for. They are not there to make the group decide. They are there to make sure the decider hears what the group knows. A CEO who speaks first gets back an echo of her own view, delivered with more confidence than anyone privately holds. The evidence Sunstein and Hastie summarize says that is not a quirk of weak teams; it is how groups behave ¹. A fair warning comes with it: some techniques decay into ritual. Their own example is the devil's advocate, which does far less than authentic dissent because everyone knows it is a game ¹. Use the moves that change who has been heard, drop the ones that only look participative, and judge a session by one test: did someone say something in it that the decider did not already know?

Before your next workshop

• Write one sentence: at the end of this session, what will exist that does not exist now?
• Name the decider and the rule for converging, and confirm both with the decider in advance.
• Agree with the most senior person when they will speak. The answer is last.
• Plan at least one moment of silent individual writing before every discussion that matters.
• Tell each participant what information they are there to bring.
• Book the follow-up: the written decisions go to attendees within two working days.

Then look again at juror 4. He had marked the highest confidence on his private ballot, and he gave it up in about a second, because of when he was asked. Every leadership team has a juror 4. The facilitator's whole job is to ask him before the others answer.

Questions

What does a strategy workshop facilitator do?

A strategy workshop facilitator designs and runs the session so that the group pools what its members know and produces the agreed output: options, a decision or a plan. In practice that means setting the question and decision rule at the start, keeping views independent until they are written down, steering the convergence, and making sure decisions are recorded and sent out afterward.

Why should the most senior person speak last in a workshop?

Because people learn from and defer to whoever speaks first, especially when that person has power. Sunstein and Hastie describe cascades in which group members follow early statements, and they advise leaders to avoid taking a firm position at the outset so that information others hold can surface. Speaking last lets the decider hear views before they bend toward the boss's.

Should the strategy team facilitate its own leadership offsite?

It can when it has no recommendation on the table. When the strategy team owns one of the options, it should present that option and let someone without a stake run the room, such as an executive from an unaffected function or an outside facilitator. This is practitioner judgment, not a research finding, but a facilitator with a stake rarely seems neutral to the room.

Is consensus a good way to end a strategy workshop?

On hard choices, consensus used as a default tends to produce wording everyone can accept and nobody can act on. Agree the rule for converging before the session: a named decider who hears the room, a vote the decider can overrule with a stated reason, or consent unless someone raises an objection framed as a risk. Any of these works if it was agreed in advance.

Where to go next

• Workshop Design Canvas: Plan Who Hears What, and When
• Getting a Room to Decide: A Decision-Making Framework
• Strategy Offsite Agenda Builder
• Pre-Mortem Worksheet
• Lencioni's Five Dysfunctions