Pyramid principle presentations: lead with the answer, then prove it

The structure: answer, arguments, evidence

Barbara Minto developed the Pyramid Principle at McKinsey, where she was hired in the 1960s and later trained its consultants to write. She set it out in The Pyramid Principle: Logic in Writing and Thinking. The idea is that any clear piece of business communication can be drawn as a pyramid with three layers.

1. The governing thought at the top: your single main point, recommendation or answer.
2. The key line beneath it: usually three to five arguments that, together, justify the governing thought.
3. The support beneath each argument: the facts, analysis and data that prove it, which can branch further down as many levels as the subject needs.

Each level answers the question the level above raises. State the governing thought and the reader asks “why?” or “how?”; the key line answers. State a key line point and the reader asks again; the support answers. Minto described this as a question-and-answer dialogue running down the pyramid. If a point below does not answer a question the point above provokes, it is in the wrong place.

The pyramid also needs an entry point. Minto's introduction (Situation, Complication, Question) leads the reader to the question that the governing thought answers. That opening is covered in our guide to the SCQA framework; this guide concentrates on the pyramid beneath it.

The three rules

Minto's rules are short, and almost every structural problem in a deck breaks one of them.

1. Ideas at any level must summarize the ideas grouped below them. The governing thought is the summary of the key line; each key line point is the summary of its support. If you cannot write a sentence that summarizes a group, the group is not yet a group.
2. Ideas in each grouping must be the same kind of idea. Three reasons, three steps or three problems can form a group. Two reasons and a step cannot. A useful test is whether you can name the group with a single plural noun: reasons, risks, phases, options.
3. Ideas in each grouping must be in logical order. Minto names four orders: deductive (an argument), chronological (time: first, second, third step), structural (the parts of a whole: north, central, south regions) and comparative (degree: most to least important, largest to smallest).

If you cannot say why point two comes before point three, the group has no order yet, and the reader will feel it.

Deductive and inductive reasoning

A grouping can hang together in one of two ways.

• Inductive: several separate points of the same kind that each support the conclusion independently. “Close the plant: it loses money, its equipment is at end of life, and the other two plants can absorb its volume.” Remove one point and the conclusion is weaker but still stands.
• Deductive: a chain in which each step depends on the one before. “Demand for the product is falling; this plant makes only that product; so the plant will run further below capacity each year; so close it.” Remove a step and the chain breaks.

Minto advised using inductive groupings at the upper levels of a document wherever possible. A deductive chain makes the reader hold every step in mind before reaching the point, and if they reject one premise they reject everything after it. An inductive key line lets the reader agree with each reason separately. Deductive logic is still useful lower down, inside a single argument, where the chain is short.

MECE grouping

MECE stands for mutually exclusive, collectively exhaustive, a term closely associated with Minto and McKinsey. Applied to a group of ideas, it means two things.

• Mutually exclusive: the points do not overlap. “Costs are too high” and “Labor costs are too high” overlap, so the reader cannot tell which point an argument belongs to.
• Collectively exhaustive: together, the points cover everything that matters to the conclusion above them. If a skeptic can name a reason you left out, the group is not complete.

MECE is easiest to achieve by splitting along a dimension that is complete by definition: revenue and cost, price and volume, existing and new customers, before and after a date. In practice, “exhaustive” means exhaustive for the decision at hand, not every conceivable factor.

Turning a pyramid into a deck

A finished pyramid maps onto a slide deck almost directly.

1. The governing thought becomes the action title of the executive summary, and the key line points become the summary's three to five bolded statements.
2. Each key line point becomes a section, opened by a divider or tracker that repeats the point as a sentence.
3. Each supporting point becomes one slide, with the point as its action title (see our guide to action titles).
4. The data beneath each supporting point becomes the slide body: one chart, table or framework that proves the title.
5. Lower levels that most readers will not need go to the appendix, ready for questions.

Because each layer summarizes the one below, the deck can be read at three depths: the executive summary alone, the section dividers, or every slide. A reader who stops at any depth still has the argument.

A worked example

A specialty retailer with 62 stores asks what to do about a group of weak locations. The team's pyramid looks like this. The numbers are illustrative.

• Governing thought: Close the 14 smallest stores over the next 12 months and put the savings into online fulfillment.
• Key line 1: The 14 stores lose $5.1M a year and have no realistic route back to profit.
• Key line 2: Most of their customers will keep buying from us online or at a nearby store.
• Key line 3: The savings fund the fulfillment capacity the online business needs anyway.

The key line is inductive: three independent reasons, ordered by degree from the point the board will challenge first to the point that makes the move attractive. Each then gets two or three supporting points. Read as slide titles, the deck runs like this:

1. Close the 14 smallest stores within 12 months and invest the savings in online fulfillment
2. The 14 stores lost $5.1M last year, and losses have grown three years running
3. Footfall at these sites has fallen 30% since 2022, twice the chain average
4. Lease terms and local demand leave no plausible plan that returns them to profit
5. Two-thirds of their customers already buy online or live within 20 minutes of another store
6. Comparable closures elsewhere in the chain kept 60% of customer spend
7. Closing the stores frees $4.3M a year, enough to fund a second fulfillment center
8. Online orders are growing 25% a year and current capacity runs out in 2027
9. Next step: approve the closure schedule and the fulfillment center business case in March

Slide 1 is the governing thought. Slides 2, 5 and 7 open each section with its key line point; the slides after each one carry the support. Read alone, the titles make the full case.

Common mistakes

• Building bottom-up and presenting bottom-up. Analysis is done from the data upward, but the deck should be presented from the answer down.
• A governing thought that is a topic: “Store portfolio review” instead of a position the audience can accept or reject.
• Mixed groups, such as two reasons and a next step under the same heading.
• Overlapping key line points, so the same evidence appears in two sections.
• Long deductive chains at the top, which lose the reader before the conclusion.
• Too many key line points. Beyond five, readers stop holding them in mind; group them into fewer, higher-level points.
• Groups with no order, where the sequence of slides could be shuffled without anyone noticing.

Pyramid principle checklist

• Can you state the governing thought in one sentence that takes a position?
• Does it answer the question your introduction raises?
• Does each key line point answer “why?” or “how?” about the governing thought?
• Is each group made of one kind of idea, nameable with a single plural noun?
• Is each group in a deliberate order: time, structure or degree?
• Are the points mutually exclusive and, for this decision, collectively exhaustive?
• Does each slide title summarize the evidence on that slide?
• Do the executive summary, the section dividers and the slide titles each make sense read alone?