Getting a Room to Decide: A Decision-Making Framework

Version seven of the same deck

Picture a packaging manufacturer, a composite most leadership teams will recognize. It is the third meeting about its oldest plant. The deck is on version seven. The finance view says close it; operations says modernize; sales worries about two big customers. Ninety minutes of good discussion later, the CEO says what she said last month: “Let’s get a bit more data and come back to this.” Everyone nods. Everyone is relieved. Nothing has been decided, and the plant keeps running the way it did last quarter.

That meeting did not fail for lack of information. It failed because nobody could answer four questions. What exactly are we deciding? What are the real alternatives? How will we judge them? Who makes the call? More data cannot answer any of those, which is why the next meeting will end the same way.

A decision-making framework is an agreed sequence for getting a group to a decision: frame the question, build options that differ in kind, agree criteria before anyone scores, name one decider, test the recommendation, get explicit commitment, and write it down. Do all of it for the few decisions that are hard to reverse, and almost none for the rest.

Bain & Company’s Paul Rogers and Marcia Blenko put a number on how rare this is: a Bain survey of more than 350 global organizations suggested that only about 15% of companies practice effective decision making ². The thesis of this page is blunter. When a room keeps asking for more analysis, it is almost never short of data. It is short of structure, and structure is cheap.

“Let’s get more data” is how a room says it doesn’t know who decides.

The framework in one table

None of these steps is new. The value is in doing all seven, in order, every time the decision is consequential. Skip one and the meeting tends to end with a request for more analysis, because the missing step is invisible and data is the only thing anyone knows how to ask for.

Frame it, and decide how much process it deserves

Write the decision as a single question. “What should we do about Europe?” is a topic. “Should we close our German sales office and serve the region through distributors from January?” is a decision. Add the deadline and the name of the person who will decide.

Then classify it. In his 2015 letter to shareholders, Jeff Bezos split decisions into two types. Type 1 decisions are consequential and irreversible or nearly irreversible, one-way doors that must be made methodically, carefully and slowly. Type 2 decisions are changeable and reversible, two-way doors, and can and should be made quickly by high-judgment individuals or small groups ⁴. His warning is that as organizations grow they tend to use the heavy Type 1 process on most decisions, including many Type 2 ones, and the result is slowness, unthoughtful risk aversion and too little experimentation ⁴.

He adds a footnote that is easy to miss and worth more than the main text. The opposite mistake, using the light Type 2 process for Type 1 decisions, is less often seen for a grim reason: companies that habitually do it “go extinct before they get large” ⁴. Both errors are real. Big companies make the first one; the second one kills you.

The rest of this framework is for Type 1 decisions. For Type 2, name an owner, set a short deadline and get out of the way. The decision stakes classifier is a quick check if you are unsure which you have.

Options that differ in kind, criteria set before anyone scores

A recommendation with no real alternatives is a request for approval. Ask for at least three options, including the status quo, and make them differ in kind, not degree. “Close the office,” “keep it and cut costs” and “sell the regional business to a partner” are three options. “Cut 10%, 15% or 20%” is one option at three settings.

Stress-test the leading option before the meeting with a premortem. In Gary Klein’s method, the team assumes the project has already failed and generates plausible reasons why, which makes it safe for knowledgeable doubters to speak up during planning ⁶. A pre-mortem worksheet keeps the exercise to about half an hour.

Agree the criteria and their weights with the decider before the options are scored. The order matters more than it looks. If the criteria arrive after the options, people pick the criteria that favor the option they already like, and they do it in good faith. Four to six criteria are enough: three-year operating profit, one-time cost, risk to key customers, time to implement, fit with the strategy. Then write down the trade-off the criteria imply. If speed and cost conflict, say which wins.

Finally, set the bar for information. Bezos’s 2016 letter suggests most decisions should probably be made with somewhere around 70% of the information you wish you had, and that waiting for 90% usually means being slow ⁵. It is a rule of thumb, not a measurement. Its real use is the question it lets you ask out loud: what would the missing information actually change?

Settle who decides

Rogers and Blenko argued that decisions stall where accountability is ambiguous, usually at one of four bottlenecks: global versus local, center versus business unit, function versus function, and inside versus outside partners ¹². Their remedy is RAPID, which names five roles: recommend, agree, perform, input and decide ². The letters are not a sequence; the authors say they took some liberties to make a useful acronym ².

The definitions follow Bain’s own description ³. Two details prevent most of the trouble. First, a veto has a price: anyone who exercises one must offer an alternative or escalate to the person with the D. Rogers and Blenko illustrate it with Wyeth Pharmaceuticals, whose top executives kept veto power over many important proposals when the company used RAPID to move faster on its drug Enbrel ². Second, if a group must hold the D, decide up front how it settles disagreement, by majority vote or by giving the chair the final say ³.

If your organization runs on RACI charts, be careful: RACI was built to assign work, and its Accountable role does not reliably mean the person entitled to choose. Our comparison of RACI, RAPID and DACI covers which model fits which fight. And before you assign roles, map who actually holds power over the outcome; stakeholder mapping is the tool for that.

Run the meeting

Send the recommendation as a written narrative memo ahead of time, or give the room silent time at the start to go through it. The meeting exists to test the recommendation, not to hear it presented.

1. Restate the question and the deciderThe chair opens with the one-sentence question, the deadline and who holds the D. If anyone disputes these, settle that first.
2. Confirm the criteriaShow the agreed criteria and weights. Any change now must be explicit and approved by the decider.
3. Test the recommendationQuestions go to the recommender. Input holders state their strongest concern; those with an agree role say whether their requirements are met.
4. Surface dissentAsk each person what would make this fail and what they would choose instead. Hear the most junior people first.
5. Decide or define the gapThe decider states the decision, or names the specific information missing, who will get it and the date of the next meeting.
6. Ask for commitmentAsk each person, by name, whether they will support the decision in their own unit. Note anyone who will not, and why.
7. Write it down before people leaveRepeat back the decision, the reasoning, the owner and the review trigger, and get agreement on the wording.

Disagree and commit, or escalate

Commitment after disagreement is what turns a decision into action. Bezos’s 2016 letter offers a phrase for it: “If you have conviction on a particular direction even though there’s no consensus, it’s helpful to say, ‘Look, I know we disagree on this but will you gamble with me on it? Disagree and commit?’” ⁵

The part most people forget is that it runs upward. He describes greenlighting an Amazon Studios original he doubted and writing back to the team, “I disagree and commit and hope it becomes the most watched thing we’ve ever made” ⁵. What he means, in his words, is “a genuine disagreement of opinion, a candid expression of my view, a chance for the team to weigh my view, and a quick, sincere commitment to go their way” ⁵. A boss who only ever asks others to disagree and commit is using the phrase as a gavel.

And it has a limit. When teams have different objectives and fundamentally different views, Bezos writes, no amount of discussion will resolve it, and without escalation “the default dispute resolution mechanism for this scenario is exhaustion. Whoever has more stamina carries the decision” ⁵. His example is Amazon’s choice to let third-party sellers compete directly on its own product detail pages, a big decision that set up hundreds of smaller ones, many of which had to be escalated to the senior team ⁵. His verdict on the alternative: “‘You’ve worn me down’ is an awful decision-making process” ⁵.

The fair objection: process is how big companies get slow

The strongest objection to everything above is that it is exactly the machinery that makes large companies slow. Roles, criteria, memos, logs: a fast company just decides. Bezos himself is the best witness for the prosecution, since his whole argument is that heavy process applied to light decisions produces slowness and too little experimentation ⁴.

The answer is in the sorting. The framework is deliberately heavy and deliberately rare. A leadership team might face a handful of true one-way doors in a year; those deserve a week of structure. Everything else should get an owner and a deadline. The companies that are slow are usually not the ones with too much structure on big decisions. They are the ones with no rule for telling big from small, so everything gets the same three meetings.

There is a second, quieter objection: the framework cannot fix the person at the top. Rogers and Blenko concede it: an indecisive decision maker can ruin any good system ². True. What the framework does is make the indecision visible. When the question, options, criteria and roles are all on paper and the decision still does not come, everyone in the room can see where the bottleneck sits. That visibility is uncomfortable, and it is the point.

Write the decision down

A decision that is not written down will be remembered differently by everyone in the room within a month. Keep a decision log with one entry per Type 1 decision, written the same day.

The assumptions and the review trigger matter most later. They let you judge the decision by what was knowable when it was made, not by how it turned out, and they tell the team in advance what evidence would justify reopening it.

1. On Monday, list the decisions your leadership team has discussed more than twice without deciding.
2. For each, write the one-sentence question and name the person who holds the D. If you cannot name one, you have found the problem.
3. Classify each as Type 1 or Type 2, and hand the Type 2 decisions to an owner with a deadline this week.
4. For the Type 1 decisions, agree the criteria with the decider before anyone builds another option.
5. Start the decision log with whatever you decide next.

The plant meeting in the opening had everything a good decision needs except the four things that make one possible. Bring those, and the data you already have is usually enough.

Questions

What is a decision making framework?

A decision making framework is an agreed sequence for reaching a decision as a group: frame the question, build real options, agree criteria, assign decision rights, test the recommendation in a meeting, commit, and write the decision down. Its purpose is to make sure the group knows what it is deciding, how it will judge the options, and who makes the final call.

What is the difference between Type 1 and Type 2 decisions?

The terms come from Jeff Bezos’s 2015 letter to Amazon shareholders. Type 1 decisions are consequential and irreversible or nearly irreversible, one-way doors that deserve a slow, careful process. Type 2 decisions are reversible two-way doors that individuals or small groups should make quickly. He warned that large organizations tend to use the heavy process for both.

What does disagree and commit mean?

It means stating your disagreement candidly, then committing fully to the decision once it is made. Jeff Bezos described it in his 2016 shareholder letter as a way to save time when there is conviction but no consensus, and said it applies to the boss too. It is not meant for deep misalignment of objectives, which he said should be escalated quickly.

Who should make the final decision in a leadership team?

One named person wherever possible. Bain’s description of its RAPID tool says there should ideally be one decider per decision, and that if a group must decide it should agree in advance how, such as by majority vote or by giving the chair the final say. In the meeting itself, it helps if the decider speaks last.

Where to go next

• RACI vs. RAPID vs. DACI: Which Decision Model, When
• The Narrative Memo: Template and Method
• Decision Log
• Decision Stakes Classifier (free question set)
• Pre-Mortem Worksheet (free worksheet)