The Pyramid Principle Literary Notes

The Pyramid Principle by Barbara Minto.


Always give the summarizing idea first before the individual ideas being summarized. Summarizing is explaining the links between the ideas.

Readers can only take sentences one at a time. The right sequencing is vital.

Vertical relationships are formed by making a statement and then answering the next logical question. E.g. pigs are great -> why, reasons a b c. You shouldn’t answer a question the reader won’t ask.

Horizontal relationships are formed by an inductive or deductive argument. A deductive argument is a statement about a situation and the second idea comments on the subject and predicate of the first idea and the third idea states the implications e.g. a is b, c is a, therefore c is b. An inductive argument is where all the ideas can be described with one plural known e.g. reasons, steps, problems.


An introduction sets expectations of what the reader knows or will know and the initial key question that it raises which will be answered in the document. This is the root node for the rest of the document. Situation, complication, question, answer SCQA.

Never write about categories, only ideas. For example: a section called ‘background’ is not helpful.

Introductions remind rather than inform. The length should be based on the needs of the reader, history background prior works, but interwoven in narrative form.

Intros as narratives build credibility and say up front what is in the supporting body.

Sometimes the S, C, A are in different order, but they are always present. Sometimes the question is implied.

There should only ever be one question (stated or implied) the document answers.


To transition between sections or chapters, use backward references. Restate the main idea from the section and state the next idea which raises the question the next section answers. You don’t need a full narrative to connect them because that was already done in the intro.

Long sections should be summarized as a form of good manners.

Conclusions are not necessary, but if you do add them do not just summarize, leave the reader with an appropriate emotion (e.g. the implication or philosophical point) or next steps that are not controversial.

When structuring the connections of key lines use inductive arguments rather than deductive otherwise there can be a large separation between the problem and the answer. Deductive arguments are best when the arguments are close together e.g. in the same paragraph.

Inductive arguments that have a single supporting argument should actually be deductive arguments otherwise it’s implied an implied connection and unclear.


Headings should also follow a pyramid pattern. There should always be more than one subheading and they should form an inductive argument supporting the parent heading. Sections (between major heading) have an ‘invisible’ fence that separates them for the reader so emphasize parallel ideas not between groups of subsections.

Headings should be concise and a reminder of the idea not dominate. Headings are more for the eye than the mind so don’t use regard them as text–the doc should flow smoothly even if you removed the headings. The opening sentence should connect the ideas between sections and not rely on the heading.

Always introduce subheadings otherwise it’s up to the reader to figure out how they connect. Subsections should never begin immediately after the heading.

A good test of headings is putting them in the table of contents to see if it fits together as an overview of the contents.

There are 4 rules to reviewing writing: you can question the order of ideas, the source in your problem solving, summary statements, and prose used to express ideas.

Impose a structure on the order of ideas to determine if it makes sense (main point is supported, lists are in a logical order e.g. structure, process or ranking, are mutually exclusive and collectively exhausted).

Problem solving

A problem is either a result you don’t like or a result you can’t explain.

Problem solving is composed of:

  • What is the problem
  • Where does it lie
  • Why does it exist
  • What could we do about it
  • What should we do about it

Another way of saying that is you must identify the gap of where you are and where you want to be, the situation that gives rise to to the gap, the underlying processes, alternatives, and a recommendation.

You’ll know when you have researched enough about where the problem lies when you have identified all the parts in the system, can arrange them in sequential order and you know the inputs and outputs. This means you understand the relationships between all of the things in the system.

What could we do about it enumerates all the ways to solve the problem.

Very often you are writing for a reader who doesn’t know the problem so it must be explained in detail.

Use Logic Trees to organize the analyses that must be performed. Typical logic trees are financial structure (ROI), task structure (increase earnings), choice structure (bifurcate by activity), and sequential structure (combines activity and sequence).

Abduction is the process of problem solving.

Summary statements

Inductive argument summaries are either the effect of the supporting actions or inference from the supporting situation statements. For example, in a list of steps the heading is the effect of each cause listed below.

Deductive argument summaries are the conclusion of the deductive statements.

Action ideas can not be grouped by similarity only by effect. Situation ideas can be grouped by similarity.

Sometimes ideas are actions disguised as statements. In that case, think about why you chose those points and what brings them together. Often it’s an indication that it’s a set of actions that lead to an effect.

Putting it into readable words

Use mental imagery to help the reader call to mind an image as they read it. This helps with recall and enjoyment. Think of it as a skeletal model you are building in the reader’s mind. To do that, visualize the relationships between ideas then use visual sounding pros e.g. feed information, deploy people.

Problem Solving in Structureless situations

In analytical reasoning, deduction, induction, and abduction all have a case, followed by a result, followed by a rule, repeating as needed. Where you start in this cycle determines the form of analytical thinking (e.g. deduction starts with a rule, induction starts with a case, abduction starts with a result).

In scientific reasoning, it’s abduction but we don’t know the underlying structure that produces the result. To find that, we create hypotheses and run experiments to prove them until the result is sufficiently explained by the rule and the underlying structure it reveals.