Reference

Frameworks: the intellectual foundation of executive communication

Part of the Executive Communication skill · loaded on demand from SKILL.md

Read this when you need the original method behind a rule in SKILL.md, when a client or partner challenges the structure, or when the standard procedure does not fit and you need to reason from first principles.

Contents

  1. Minto's Pyramid Principle
  2. SCR and the storyline
  3. Aristotle: the three means of persuasion and arrangement
  4. McKinsey presentation practice: elevator test, one message per chart, pre-wiring
  5. The ghost deck and the dot-dash storyline
  6. How the frameworks combine in a consulting deliverable

1. Minto's Pyramid Principle

Origin and editions

Barbara Minto joined McKinsey in the 1960s and was asked to improve the firm's written reports. The method she developed was first taught inside the firm, then circulated as a manual and published as "The Pyramid Principle: Logic in Writing and Thinking" (Pitman edition, 1987; earlier printings exist from the late 1970s), revised and expanded as "The Minto Pyramid Principle" (1996), with further editions through 2009. It is the standard writing text of the major strategy firms. Minto also coined the term MECE (mutually exclusive, collectively exhaustive), which the pyramid depends on.

The claim

A reader can absorb a set of ideas only if they arrive pre-organised: one idea at the top, supported by groups of ideas beneath it, each group summarised by the idea above. Presented any other way (in the order the work was done, in the order the writer thought of things, as a list of findings) the reader must build the structure for themselves, and will either build a different one from the writer's or stop reading.

The mechanism

Minto rests the argument on a limit of working memory that George Miller described in 1956: people can hold roughly seven items at once, and fewer when the items are unrelated. A reader who receives facts before the point must hold each fact while waiting to learn what it is for. Once the point arrives, the reader must reconstruct the connection between every fact and the point. If the point arrives first, each fact is filed as it arrives, and the reader's memory holds one structure rather than a pile. This is why "answer first" is not a stylistic preference but a constraint on comprehension, and why it applies to a paragraph, a memo, a deck and a spoken briefing alike.

The three rules of the pyramid

Minto states three rules that every pyramid must obey. She presents them as the tests a writer runs on a draft, and they are the tests a partner runs on a storyline.

  1. Ideas at any level must be summaries of the ideas grouped below them. The summary states the point the group makes, not the fact that a group exists. "Three factors drove the decline" is not a summary; "The decline is a pricing problem" is.
  2. Ideas in each grouping must be the same kind of idea. A group cannot contain two causes and one action, or two findings and one recommendation. If the members are different kinds, the group cannot be summarised, which is why rule 1 fails whenever rule 2 does.
  3. Ideas in each grouping must be logically ordered. Minto allows four orders, which correspond to the way the group was formed: deductive order (an argument), chronological order (a process), structural order (a division of a whole, such as regions or functions), and comparative order (a ranking by degree or importance). The writer must be able to say which order the group is in and why.

The vertical relationship: question and answer

The pyramid is read top-down. Each idea, once stated, raises a question in the reader's mind ("Why?", "How?", "How do you know?"), and the ideas at the next level down must answer that question. The vertical relationship therefore enforces relevance: an idea that does not answer the question raised by the idea above it does not belong there. A writer checks the vertical relationship by writing down the reader's question after every summary sentence and confirming the group below answers exactly that question. The most common failure is a group that answers a different question from the one raised, usually because the writer wanted to include an analysis that was expensive to produce.

The horizontal relationship: deductive or inductive

Ideas within a group relate to each other in one of two ways.

Deductive: a chain of reasoning. A major premise about the world, a minor premise that comments on the major premise (its subject or its predicate), and a conclusion that follows from both. "Margin fell because of price. Price fell because of unmanaged discounting. Therefore fix discounting." The summary of a deductive group is the conclusion, usually with the premises compressed into it ("Fix discounting, because that is where the margin went"). Deduction is powerful because the reader who accepts the premises must accept the conclusion, and dangerous for the same reason: a reader who rejects one premise rejects everything after it. Minto advises keeping deductive chains short (three or four steps) and warns against using deduction at the key line for a sceptical audience, because it forces the reader to accept the whole chain before hearing the recommendation.

Inductive: a set of ideas of the same kind, from which a general point is inferred. "Discounts below list doubled. Realised prices for the same SKU vary 30%. Overrides are approved without margin review." The summary is what these have in common or imply: "Pricing is unmanaged." The test of an inductive group is that all members can be described by one plural noun ("three symptoms of unmanaged pricing") and that the summary is an inference from them, not a count of them. Induction is more forgiving: if one member is disputed, the others still support the summary. It is weaker at compelling action, because a list of observations does not by itself say what to do.

Minto's practical rule is that the two are combined: the top of a recommendation pyramid is usually a short deductive argument (situation, cause, therefore action), and each premise of that argument is supported inductively by evidence.

The introduction: Situation, Complication, Question

The pyramid starts with an answer, but the reader needs a reason to want one. Minto's introduction supplies it as a short narrative:

  • Situation: a statement about the subject that the reader already knows and accepts. Its job is to establish common ground and time and place. It must not contain anything the reader would dispute.
  • Complication: what happened within the situation that created a problem, a change, or a need. It is the tension that makes the reader want the answer.
  • Question: the question the complication raises, which the governing thought at the top of the pyramid answers.

Later editions add the Answer as the fourth element (SCQA). Minto's guidance on length is that the introduction should be as short as the reader's existing knowledge allows: for a reader who lived through the complication, two sentences; for a board reading cold, a paragraph. The Situation and Complication are the writer's chance to show the reader that the problem has been understood before the answer is given, which is also where ethos is first established.

Minto also describes standard question patterns that the complication typically produces: something has gone wrong (what do we do?), something might go wrong (how do we prevent it?), we know what to do (how do we do it?), we did what we planned and it did not work (why not?), and we have several options (which one?). Knowing which pattern the engagement is in tells you what kind of governing thought is required.

Summarising a group: action ideas and situation ideas

Minto's most practical and least followed instruction concerns how to write the sentence above a group. She distinguishes two kinds of idea that can be grouped.

Action ideas tell the reader to do something (steps, recommendations, processes). The summary of a group of action ideas states the effect of carrying them all out: not "take these three steps" but "these three steps recover 3.9 points of gross margin". The test is to ask what the reader has once every action in the group has been done; that result is the summary. Actions in a group must also be at the same level of abstraction, so that "renegotiate twelve accounts" does not sit beside "improve profitability".

Situation ideas tell the reader something is the case (findings, facts, observations). The summary of a group of situation ideas states what their similarity implies: not "there are three findings about pricing" but "pricing is unmanaged". The procedure is to find the word or structure the members share (all are about discounts, or all are about the same customers), and then ask what that shared feature means for the reader.

Failing to do this produces what Minto calls the intellectually blank assertion: a summary that announces a group without saying what it means. "Three factors explain the decline", "several issues were identified", "the analysis covered four areas". Each of these leaves the reader to do the summarising, which is the work the writer was supposed to do. A partner reading a storyline circles every such sentence and returns the draft.

The four logical orders

Rule three requires that ideas in a group be ordered, and Minto ties each order to the mental process that produced the group. Naming the order is a check that the group is genuine.

OrderMental processExample in a storylineTest
DeductiveReasoning: premise, premise, conclusionMargin is in gross margin; gross margin is price and mix; therefore fix price and mixEach step comments on the one before
ChronologicalDetermining cause and effect, or a sequence of stepsWave 1, wave 2, wave 3The steps happen in that order
StructuralDividing a whole into partsFourteen branches; four product lines; three regionsThe parts add to the whole, no overlap
Comparative (degree)Ranking by importance or sizePrice realisation 1.3 points, mix 0.6, inflation 0.2The reader can see why the first is first

Two common errors: a group with no discernible order (which usually means it is not a group but a collection), and a structural group that is not exhaustive (three of four regions), which the reader notices and which undermines everything above it.

The introduction's question patterns

Minto observes that complications fall into a small number of patterns, each of which produces a predictable question and therefore a predictable kind of governing thought.

ComplicationQuestionKind of answer needed
Something has gone wrongWhat do we do about it?A recommendation and its rationale
Something might go wrongHow do we prevent it?A risk assessment and a plan
We know what to doHow do we do it?A plan and its sequence
We did what we planned and it did not workWhy not?A diagnosis, then a revised recommendation
We have several optionsWhich should we choose?A choice with the criteria and the trade-offs
We are not sure the recommendation is rightIs it?A test and its verdict

Identifying the pattern before writing tells the team what the governing thought must contain. The Northgate worked example is "we did what we planned (a 2023 price rise) and it did not work, and we are now being asked to do something else (cost)", which is why the storyline is diagnostic before it is prescriptive.

The key line

The key line is the set of ideas immediately beneath the governing thought. It is the most important level in the document because most senior readers read only the governing thought and the key line. Minto recommends stating the key line in the introduction, immediately after the answer, so the reader knows the shape of the whole argument before reading any support. In a deck this becomes the executive summary; in a memo it is the second paragraph.

Building the pyramid: top-down and bottom-up

Minto gives two procedures. Top-down: write the Situation and Complication, derive the Question, write the Answer, then ask what question the Answer raises and answer it at the key line, and so on downward. Use it when the answer is known. Bottom-up: list every point you want to make, work out the relationships between them, group them, draw the conclusion each group supports, then build upward until one governing thought remains. Use it when the material exists but the answer is unclear. In practice a consulting storyline uses both: the ghost deck is top-down (a guessed answer directs the work), and the final storyline is checked bottom-up (does the evidence actually support what the titles claim?).

Failures the pyramid prevents

  • The narrative report: the work in the order it was done, with the conclusion at the end.
  • The topic outline: headings that name subjects ("Market", "Costs", "Options") and so cannot be summarised.
  • The intellectually blank summary: "There are three issues" instead of what the three issues mean.
  • The mixed group: findings and recommendations in the same list.
  • The unanswered question: a summary that raises "why?" followed by support that answers "how?".

2. SCR and the storyline

Situation, Complication, Resolution is the form Minto's introduction takes in consulting storylines and executive summaries. It has no separately published author; it is a working convention, described in practitioner literature and in the public accounts of the strategy firms, in which the Question is left implicit and the Resolution (the answer) is stated directly after the Complication.

SCR works for the same reason a story works: the Complication introduces tension and the Resolution relieves it. A document with no Complication gives the reader no reason to want the Resolution; the most common weak executive summary states a Situation and then a list of findings, and the reader does not know why any of it matters.

SCR is used at three scales. At document scale it is the executive summary. At section scale each key-line point can be introduced with its own short SCR ("Costs looked like the obvious lever [S]; but opex is at peer median [C]; so the answer lies in gross margin [R]"). At slide scale the action title is often the R, with the S and C in the transition from the previous slide. A storyline that reads as a chain of small SCRs is one the audience can follow without effort.

One caution: the Complication must be real and specific, not manufactured drama. A Complication the reader does not recognise ("The industry is at an inflection point") costs credibility. A Complication the reader recognises immediately ("The board asked for a cost programme, and costs are not where the margin went") earns it.

3. Aristotle: the three means of persuasion and arrangement

Aristotle's "Rhetoric" (fourth century BC) is the earliest systematic treatment of persuasion and remains the source of the vocabulary. Two parts of it matter for consulting deliverables.

The three means of persuasion (Rhetoric I.2)

Aristotle distinguishes the means of persuasion the speaker controls from those that exist independently (witnesses, documents). The three technical means are:

  • Ethos, the character of the speaker as it appears in the speech. Persuasion happens when the speaker seems credible, and Aristotle is specific that this credibility must be produced by the speech itself, not assumed from reputation. It rests on three things the audience judges: practical competence, good character, and goodwill toward the audience. For a consultant this means that the deck must display competence (the method is visible, the numbers reconcile, the limits are stated), honesty (what the team did not find is reported as clearly as what it did), and goodwill (the recommendation is in the client's interest, and the hard message is delivered with care for the people it affects).
  • Pathos, the disposition of the hearer. Judgment is affected by the state the audience is in, and the speaker can shape that state. Aristotle treats the emotions analytically: each has conditions under which it arises, and a speaker who establishes those conditions produces the emotion without naming it. For a consultant this means the audience must feel the cost of inaction and the attraction of the prize, produced by concrete cases and numbers, never by emotional vocabulary. A slide that says "urgent" produces nothing; a slide that shows the twelve largest accounts losing money produces concern.
  • Logos, the argument itself. Aristotle's two forms are the enthymeme (a deductive argument from premises the audience already accepts, usually with one premise left unstated because it is obvious) and the example (an inductive argument from particular cases). These map directly onto Minto's deductive and inductive horizontal relationships. The enthymeme is important for consultants because the strongest argument uses a premise the client already holds: if the CEO already believes the sales force discounts too freely, the argument starts from there.

Ethos, pathos and logos in a consulting deliverable, by element:

ElementEthosPathosLogos
SCQ introductionShows the problem is understood in the client's termsThe Complication makes the cost of the status quo feltSets the question the pyramid answers
Executive summaryStates method and limits in two linesLeads with the number at stakeGoverning thought and key line
Body slidesSource lines, reconciliation, sample sizesOne concrete case per key-line pointTitles as claims, exhibits as proof
Risks slideAdmits what could go wrong and what is unknownShows the mitigations are realRefutation of known objections
Decisions slideAsks for what is needed, nothing moreStates what is lost by waitingThe ask follows from the proof
Pre-wiringGoodwill: no one is surprised in publicObjections heard and absorbed in privateErrors corrected before the room

The enthymeme deserves a note because it is the form most consulting arguments take without knowing it. An enthymeme leaves one premise unstated because the audience already holds it. "Opex is at peer median, so cost is not the lever" relies on the unstated premise that a company at the peer median has no large cost opportunity. If the audience does not hold that premise (a CEO who believes the whole peer set is inefficient), the argument fails silently. The discipline is to write out the unstated premise for every key-line step and confirm the audience holds it; if not, state and prove it.

The failure this framework prevents is the belief that logos is sufficient. A decision-maker judges the messenger and the stakes as well as the logic, and an analytically correct recommendation delivered by a team the audience does not trust, about a problem the audience does not feel, is rejected.

Arrangement (Rhetoric III.13-19)

Aristotle argues that a speech has only two necessary parts: the statement of the case and the proof of it. In practice he allows a proem (introduction) before and an epilogue after, giving four parts: proem, statement, proof, epilogue. He is dismissive of elaborate subdivisions. Later Roman writers (Cicero in "De Inventione", the anonymous "Rhetorica ad Herennium", Quintilian in the "Institutio Oratoria") expanded this into six parts: exordium (introduction, winning attention and goodwill), narratio (statement of the facts), partitio or divisio (the points in dispute and the order of treatment), confirmatio (proof), refutatio (answering opposing arguments), peroratio (conclusion, summarising and moving the audience to act). They also allowed the parts to be reordered when it strengthened the case.

Mapped onto a consulting presentation:

Classical partConsulting equivalentWhere it lives
Exordium / proemSituation and Complication, plus the credentials established by the framingFirst minute; first paragraph of the summary
Narratio / statementGoverning thought and key lineExecutive summary slide
PartitioThe agenda, which is the key line as headingsOptional; often folded into the summary
Confirmatio / proofThe body slides, in pyramid orderMain deck
RefutatioThe objections answered before the decision is requestedRisks and mitigations slide, and pre-wiring
Peroratio / epilogueDecisions requested, next steps, and the cost of not decidingFinal slide

The one placement rule that matters: refutation comes before the peroration. A presentation that ends on objections ends without a decision.

The five canons of classical rhetoric (invention, arrangement, style, memory, delivery) are also a useful checklist: invention is the analysis, arrangement is the pyramid, style is the prose polish, memory is rehearsal, delivery is the spoken presentation.

4. McKinsey presentation practice

The public accounts of McKinsey's habits (Rasiel, "The McKinsey Way", 1999; Rasiel and Friga, "The McKinsey Mind", 2001) describe several rules that apply Minto's structure to the spoken and visual presentation. They are grounded in these two books; no internal firm document is cited.

  • The elevator test. Know the recommendation well enough to explain it clearly to the client in thirty seconds. The test is practical (meetings get cut, executives leave early) and diagnostic: if the team cannot pass it, the governing thought is not yet clear to the team.
  • Structure. A presentation takes the audience down the path of the logic in clear steps. This is the pyramid, spoken.
  • One message per chart. A chart makes one point; the more it tries to say the less it says. The action title states the point and the chart proves it. Charts are a means of communicating the message, not an art project.
  • Keep it simple. Simplicity in the exhibit serves the message.
  • Prewire everything. A good business presentation contains nothing new for the audience. The team walks each player at the client through the findings before they are gathered in one room, so that the meeting confirms rather than reveals. Objections are heard in private, where they can be absorbed; the recommendation that will not survive is changed before it is presented.

The reasoning behind pre-wiring is worth stating because it is often resisted as a loss of drama. A decision-maker who first sees a difficult finding in front of peers has two options, to accept it publicly or to attack it, and the second is safer for them. The same person, shown the finding alone, can ask questions, correct errors and arrive at the meeting as a supporter. Pre-wiring converts opposition into co-authorship.

5. The ghost deck and the dot-dash storyline

These are the working tools of the storyline in strategy consulting. They are practitioner conventions rather than published frameworks; the terms appear in public accounts by former consultants and in practitioner guides.

The dot-dash storyline is the pyramid written as an outline. Each dot is one message, which will become one slide's action title, written as a full sentence. Beneath each dot, dashes list the supporting points and the evidence, with a note of the exhibit that will carry them. The dot-dash is cheap to write and cheap to change, which is its purpose: the argument is debated and revised at this stage, before any slide exists. A partner reviewing a dot-dash reads the dots alone first; if they form the argument, the structure is sound and the review moves to whether the dashes prove the dots.

The ghost deck (also "ghost pages" or a "storyboard") is the deck drafted before the analysis is complete: each slide has its action title and a sketch of the exhibit, with no data. It serves three purposes. It directs the work, because every analysis must feed a slide and any that does not is cut. It aligns the team and the partner on the shape of the answer early, so that effort is not spent on slides that will not be used. It allows early pre-wiring with a skeleton the client can react to. The ghost deck is a hypothesis about what the analysis will show; when the analysis disagrees, the titles change and the findings do not. Because it requires guessing the results, ghosting is iterative and revisions after analysis are normal.

The sequence in practice: SCQ and governing thought, then dot-dash, then ghost deck, then analysis, then revised dot-dash, then final titles, then exhibits, then executive summary written last (so that it reconciles to the deck), then speaking notes, then pre-wiring, then the meeting.

6. How the frameworks combine in a consulting deliverable

LayerFrameworkWhat it governs
Why the reader wants an answerMinto's SCQ; SCRThe introduction and the executive summary's first paragraph
What the answer is and how it is supportedMinto's pyramid: governing thought, key line, three rules, vertical Q&A, horizontal logicThe storyline, titles and body order
Whether the reader believes the messenger and feels the stakesAristotle's ethos and pathosMethod transparency, stated limits, concrete cases, framing of the cost of inaction
The order of the spoken presentationAristotle's arrangement, as expanded by the Roman rhetoriciansProem, statement, proof, refutation, epilogue
The working methodGhost deck, dot-dash, one message per chart, elevator testHow the deliverable is drafted and reviewed
The meetingPrewire everythingWho is briefed, in what order, before the room

A deliverable that obeys only the pyramid is logically sound and may still fail in the room. One that obeys only rhetoric is persuasive and may be wrong. The engine requires both, in that order: logic first, because the analysis must be right; rhetoric second, because being right is not enough.