The Pyramid Principle

Anyone who works in data has written this report. First, where the data came from. Then how you cleaned it. Then a dozen charts. And on the very last slide: "So, we recommend…" Your boss flips to page three and asks, "What's the conclusion?"

This book was written for that moment.

Barbara Minto's own story is a good one. She graduated from Harvard Business School in 1963, one of about 8 women in a class of roughly 600. That same year she joined McKinsey as its first female MBA consultant. At McKinsey she noticed that when a consultant's report was hard to follow, the problem usually wasn't the prose. It was that the thinking underneath hadn't been sorted out. She turned her fix into an internal writing course, later left to work on her own, and wrote the course up as this book. Decades on, it's still required reading for new hires at many consulting firms.

Lead with the answer. Your reader's attention is limited. Give the answer first, then the reasons, then the evidence. Flip that order and the reader has to guess what you're getting at the whole way through.

Every level sums up the one below it. A piece of writing is a pyramid. At the top sits one core idea, and every point at each level has to summarize the points grouped under it. Each level answers the "why?" or "how?" raised by the one above.

Groups shouldn't overlap or leave gaps. Points at the same level should be distinct from each other and, together, complete. This is the famous MECE rule (mutually exclusive, collectively exhaustive). If two of your three reasons are really the same reason, the reader instantly senses you haven't thought it through.

Open with a tiny story. Minto suggests starting with Situation, Complication, Question, Answer: the background everyone agrees on, then what changed, then the question that change raises. Your document is the answer.

It's especially useful for data people, because the natural order of analysis (pull the data, clean it, model it, conclude) is the exact reverse of the best order for presenting it. We tend to explain what we did. The audience only wants "so what?" The pyramid forces you to answer one question before you write a word: what's the one sentence I most want them to remember? If you can't say it, the analysis isn't finished.

It has its limits. It reads like a textbook, the examples are decades-old business cases, and it can be dry. Its value isn't in the writing. It's in a habit: think clearly first, then speak clearly.

Read it if you often feel you did a mountain of analysis and still can't explain it.

Ideas at any level in the pyramid must always be summaries of the ideas grouped below them.