The Minto Pyramid Principle

What the Pyramid Principle is

You have sat through this presentation. Twenty minutes of background, methodology and context, building carefully toward a recommendation that arrives on the second-to-last slide — by which point two people are on email and the one who matters has already decided. The information was all there. The structure buried it.

Minto's fix is one sentence, and it is hers: your thinking will be easy for a reader to grasp if you present the ideas organized as a pyramid under a single point. One governing conclusion sits at the top. Directly beneath it, the handful of arguments that support it. Beneath each of those, the evidence. The reader meets the answer immediately and then descends only as far as their scepticism requires.

She developed it inside McKinsey — hired in Cleveland in 1963 as the firm's first female MBA, moved to London in 1966, left in 1973 — and published The Pyramid Principle: Logic in Writing and Thinking in 1985, expanded in 1996 as The Minto Pyramid Principle: Logic in Writing, Thinking and Problem Solving. She is also the author of MECE, the mutually-exclusive-collectively-exhaustive test that governs whether a group of arguments is a legitimate group at all. Both survived her: the firm still teaches the pyramid to new hires, and most of its competitors teach some version of it too.

When to use it — and when not

The Pyramid Principle is a format for delivering a verdict. That makes it excellent in a narrow set of situations and actively wrong in several others, which the popular summaries tend to skip.

• ✓Use it when you have a recommendation. A decision memo, a proposal, a board paper, a case answer — anything whose purpose is to get a specific answer accepted or rejected.
• ✓Use it when the audience is time-boxed and senior. People who will stop reading the moment they have what they need are exactly the people the structure is built for.
• ✓Do not use it when you have no recommendation yet. A pyramid with nothing at the apex is a fact dump wearing a suit. Interim findings want a different shape.
• ✓Do not use it when bad news needs its context first. Leading with the verdict on a failure invites the audience to litigate the verdict instead of understanding the cause.
• ✓Do not use it when the audience has to be brought to care. Conclusion-first assumes the reader already wants the answer. If they do not yet know they have a problem, you need a narrative that builds one.

Those exceptions carry more weight than they used to. Brent Dykes, who works on data storytelling, argues that the Pyramid Principle was built for executive efficiency rather than for persuasion, and that in it the answer is explicit and the insight is implicit — available to anyone who digs for it but never surfaced as the main point. He concedes what it is genuinely good at: organizing information so it can be consumed quickly and efficiently. That is the honest boundary. It is a format for informing a decision-maker, not for changing a mind that is not yet in the room. The three rules below are what separate a real pyramid from a list with a heading:

• ✓Every idea summarises the ideas grouped beneath it. If the point above is not what the points below add up to, the structure is decorative.
• ✓Ideas in a group are logically the same kind. Three causes and a recommendation is not a group. This is the MECE test doing its work.
• ✓Ideas in a group are in a deliberate order. Chronological, structural, or by importance — but chosen, and the same order maintained wherever that group reappears.

Step by step, with a worked example

Take a real decision: whether to build a billing integration in-house or buy it. The pyramid is assembled top-down, but it is discovered bottom-up — which is the part most people skip.

1. 1Write the SCQ introduction first. Situation: we process billing through a vendor integration built in 2021. Complication: the vendor is deprecating the API in nine months. Question: do we rebuild in-house or move to a new vendor? Minto's own name for this is the SCQ Framework — Situation, Complication, Question. The A that most summaries append is your pyramid; it is not part of her introduction.
2. 2Answer the question in one sentence. Move to a new vendor. That sentence is the apex. If you cannot write it yet, stop — you are not ready to write, and the pyramid has just told you so.
3. 3Group the arguments that support it. Cost over five years, time to migrate, and the engineering capacity the rebuild would consume. Three arguments of the same kind: reasons this option wins.
4. 4Apply MECE to the group. Do the three overlap? Does anything decisive sit outside them — vendor lock-in, data residency? Overlaps make the argument feel repetitive; gaps are where the objection you did not prepare for comes from.
5. 5Put the evidence under each argument. The five-year model under cost, the migration estimate under time, the roadmap under capacity. Nothing floats: every number answers to the argument above it.
6. 6Read it backwards as a question test. Each point should raise the exact question the points beneath it answer. Where it does not, you have found a broken rung — and it is always cheaper to find it here than in the meeting.

The pyramid as an argument tree

Read the rules again and the shape is familiar: a claim, supported by a small group of arguments, each supported by evidence, with a test for whether the group is legitimate. That is an argument tree drawn on paper — which is both why the method has lasted forty years and where the paper version runs out.

• ✓A pyramid has no place to put an objection. The structure holds supporting arguments; the counter-argument that nearly changed the recommendation lives in a comment thread, or in someone's memory, or nowhere.
• ✓It is authored, not deliberated. One person builds it and presents it. The disagreement happens afterwards, against a finished artifact, and leaves no trace in it.
• ✓It is a snapshot. A document written in March cannot show which of its arguments survived April. When the question reopens, the pyramid tells you what was concluded but not what withstood scrutiny.

Pyramid Principle vs the alternatives