Decision Meeting Checklist: Most of It Happens the Day Before

The factory plan that got a green light by accident

Michael Mankins opens with a hypothetical meeting at a company he calls AnyCo. The CEO's assistant collects agenda items by email, lists them in the order they arrive and guesses how long each will take. In the meeting, the team spends a long time on the look and feel of an advertising campaign and turns a compliance item into a gripe session about IT. A major capital proposal for a factory automation program gets a green light after a cursory examination, to the CFO's great discomfort. The meeting breaks up an hour late ¹.

Nobody in that room was careless. The decision was lost before the meeting began: the agenda was a list in order of arrival, no one had marked which item needed a decision, and no one had asked for alternatives. Mankins's survey of top teams at 187 large companies found that more than 65% of leadership meetings were not even called to make a decision, and only 12% of executives believed their meetings consistently produced decisions on important issues ¹.

A decision meeting is a meeting called to make one named choice, with the decider present and the options already on paper. The checklist below has three parts: a go/no-go test the day before, the questions to ask in the room, and what to write down before anyone leaves. The first part matters most. A decision meeting is mostly settled the day before. If the question, the real options and the decider are not written down by then, hold a discussion, and call it one.

If you cannot name the decider the day before, you are holding a discussion.

How to use the checklist

Use it for decisions that are consequential and hard to reverse: a capital project, an acquisition, a market exit, a reorganization. Daniel Kahneman, Dan Lovallo and Olivier Sibony make the same cut for their decision-review checklist: it is not for routine decisions an executive rubber-stamps, and its sweet spot is decisions that are both important and recurring, such as approving an R&D project, a large capital expenditure or a midsize acquisition ⁴.

Run it as a pause, not a script. Whoever owns the meeting (usually the recommender or the chief of staff) checks Part 1 the day before and sends the result to the decider. Part 2 sits in front of the chair. Part 3 is filled in during the last ten minutes. Every item is a question with a yes or no answer, because a checklist you can answer with “sort of” is a checklist nobody runs.

If the decision itself is still badly framed, start one step earlier with the decision-making framework. This page assumes you know what you are deciding and need the meeting to actually decide it.

Part 1: the day before (go or no-go)

Check each item. Any “no” in the first four means the meeting is not ready to decide. Either fix it today or change the agenda item to “discussion” and say so.

1. 1. The question is written as a choiceOne sentence, with at least two answers someone could seriously argue for. “Do we close the Lyon plant by June, or retool it?” is a question. “Lyon update” is a topic.
2. 2. The decider is named and will attendOne person. Paul Rogers and Marcia Blenko call the decider the single point of accountability, who brings the decision to closure and commits the organization to act; their first warning is to ensure only one person has the D, because two people who think they decide produce a tug-of-war ².
3. 3. There are at least three real alternativesNot one plan at three sizes. Mankins argues management needs at least three real alternatives, not minor variations on one theme, before a strategy is discussed or approved; only 14% of the executives he surveyed were consistently presented with any alternative strategies ¹.
4. 4. The pre-read went out and is marked for decisionCadbury Schweppes's executive committee distributed reading materials at least five days ahead and attached a cover sheet saying whether each item was for information, for discussion or for decision ¹. Five days is their practice, not a law; pick a lead time and keep it.
5. 5. The criteria are agreed with the deciderFour to six criteria, and which one wins when they conflict, settled before anyone scores the options.
6. 6. People with a veto are few and know itRogers and Blenko warn that a proliferation of people with veto power can paralyze recommenders ². List who must agree, for example legal or a regulator, and confirm they have seen the recommendation.
7. 7. The leading option has been attackedA premortem was run: the team imagined the project had failed and wrote down why, independently, before discussing it ³. Attach the list.
8. 8. Dissent inside the recommending team is visibleKahneman, Lovallo and Sibony suggest that an absence of dissent in a team tackling a complex problem should sound an alarm ⁴. If the recommendation arrives unanimous, the decider privately asks who argued for something else.

Part 2: in the room

The meeting exists to test the recommendation, not to hear it presented. The chair runs these in order.

1. 9. Restate the question, the decider and the deadlineThirty seconds. If anyone disputes them, settle that before anything else.
2. 10. Hear the doubters before the bossThe decider holds their view until others have spoken. Sunstein and Hastie note that leaders who state their views early promote self-censorship ⁵. Ask the most junior relevant person first.
3. 11. Ask where the numbers came fromWhich numbers are facts and which are estimates, and who put the first number on the table? These are among the questions Kahneman, Lovallo and Sibony suggest for spotting anchoring ⁴.
4. 12. Ask whether the worst case is bad enoughTheir observation is that the worst case is rarely bad enough ⁴. Ask what could happen that nobody has modeled, and how sensitive the case is to competitors' responses.
5. 13. Ask what you would want to know in a yearIf you had to make this decision again in a year, what information would you want, and can you get more of it now ⁴? If the answer is cheap to get, get it. If it is not, decide.
6. 14. Decide, or name the gapThe decider states the decision, or names the specific missing information, who will get it, and the date of the next meeting. “Let's come back to this” is not an option.
7. 15. Ask for commitment by nameEach person whose team carries out the decision says whether they will support it. Disagreement said in the room can be handled. Disagreement saved for the hallway cannot.

Part 3: before anyone leaves

Mankins found that the biggest challenge for many top teams is agreeing on what they agreed to in the meeting. Participants leave with different interpretations, and those who think the decision was wrong can exercise a silent veto by relaying nothing to their teams ¹. The cure is ten minutes and a template.

Item 18 is the one teams skip and the one Mankins stresses: several companies he studied tied strategy approval to a performance contract specifying the time, talent and money required and the results management pledged to deliver ¹. A decision with no resources attached is a wish. Keep the entries in a decision log so the next review can judge the decision by what was known at the time.

The checklist, filled in

The forecast question and the worst-case question changed the outcome. Neither needed new analysis. Both needed someone to ask them in that order, with the CEO still quiet.

When to cancel, and other judgment calls

Cancelling a decision meeting feels like failure. It is usually the most useful thing the checklist does. A meeting that fails items 1 to 4 will end, at best, in a request for more data, and the people in it will learn that decision meetings do not decide. Turn it into an honest discussion slot instead, and set the date by which the four items will be done.

• The decider will not attend. Postpone. A deputy who decides “subject to the boss's review” has not decided anything, and the room knows it.
• The decider is a committee. Agree the rule before the meeting: majority, or the chair decides after hearing the committee. Write it on the agenda.
• The decision is reversible and cheap. Do not hold the meeting at all. Give the choice to an owner with a deadline. This checklist is for decisions where being wrong is expensive.
• The recommendation has only one realistic option. Kahneman, Lovallo and Sibony warn reviewers about option sets that include only one realistic alternative, the one the recommending team prefers ⁴. Send it back for real alternatives before the meeting, not during it.

The case against checklists in the boardroom

The strongest objection is that senior judgment cannot be reduced to 22 boxes. Executives who are good at decisions will resent the ritual, and executives who are bad at them will tick the boxes and decide badly anyway. Worse, a checklist can become a shield: “we followed the process” replaces “we thought hard.”

The authors of the best-known decision checklist agree with the warning. Kahneman, Lovallo and Sibony write that the goal is not to create bureaucratic procedures or to turn decision quality control into another element of “compliance” that can be delegated to a risk assessment unit, but to stimulate discussion and debate ⁴. That is the right test for every item above. If an item has become a box someone ticks without a conversation, cut it or change the question.

But the objection misses who the checklist is for. It is not there to replace the decider's judgment. It is there for the moments when judgment is not the scarce thing: when the meeting is running late, when the recommender is senior, when the room already agrees. Those are exactly the moments the AnyCo team approved a factory with a cursory look ¹. Experienced people skip steps under pressure precisely because they are experienced. The checklist is the pressure's counterweight.

Monday: run Part 1 on your next big meeting

Take the next meeting on your calendar that is supposed to decide something. Check items 1 to 4 now. Most teams find at least one “no.” Fix it, or rename the agenda item honestly.

A decision meeting is not where the decision gets made. It is where a decision that was made ready gets made official. The work of making it ready happens the day before, on paper, and a room of busy executives cannot do it for you in the hour they have.

Questions

What is a decision meeting?

A decision meeting is a meeting called to make one named choice. The question is written as a choice between real alternatives, the decider is named and present, the pre-read has gone out marked for decision, and the meeting ends with a written decision, an owner, resources and a review trigger. Meetings that lack these are discussions, which are useful but should be called that.

What should be on a decision meeting agenda?

The question and the decider, restated; a short summary of the recommendation and alternatives (not a long presentation); questions that test the recommendation, such as where the numbers came from and whether the worst case is bad enough; the decision, or the named gap and the date to return; commitments by name; and ten minutes to write down the decision, resources, owner, dissent and review trigger.

How far in advance should a decision pre-read go out?

Long enough that people actually read it. Cadbury Schweppes's executive committee, as Michael Mankins describes it, distributed materials at least five days before each session and marked each item as for information, discussion or decision. Pick a lead time that suits your calendar, keep it consistently, and move items whose papers are late.

When should you cancel a decision meeting?

Cancel or convert it to a discussion when the question is not written as a choice, the decider is not named or cannot attend, there are not at least three real alternatives, or the pre-read has not gone out. A meeting missing those will usually end in a request for more data. Set a date by which they will be fixed instead.

Where to go next

• Getting a Room to Decide: A Decision-Making Framework
• Strategy Workshop Facilitation: Who Speaks First Decides
• Decision Log
• Pre-Mortem Worksheet
• The One-Page Decision Memo (free PDF)