Pyramid Principle

Quick facts

Mechanism: Produces a pyramid that structures a recommendation answer-first - a single governing thought on top, a small set of MECE, deliberately ordered key arguments beneath it, and supporting evidence under each (Minto), with an optional SCQA intro framing.

Produces: pyramid

Evidence: P · Domain: Synthesis

Use when a conclusion or recommendation already exists and must be communicated clearly to a busy or senior reader, when findings need to be ordered into a tight top-down case, or when a write-up buries its point under context.

Most recommendations are delivered in the order they were discovered - context, then analysis, then, eventually, the point - forcing the reader to hold a pile of facts in mind and guess where they lead. The pyramid principle inverts that: the conclusion (the governing thought, the one thing the reader should do or believe) goes first; beneath it sit a small set of MECE key arguments that together justify it; beneath each sits its supporting evidence. The reader descends only as far as their trust requires and can stop at the top with the recommendation in hand. The output is a structured pyramid (governing thought + ordered key lines + support), not a flowing essay. It composes a clear case; it does not check whether the case is sound.

When to Use

• A conclusion or recommendation already exists and must be communicated to a busy or senior reader who needs the headline first.
• Findings, analysis, or a memo need to be ordered into a tight top-down case instead of a chronological narrative.
• A draft buries its point under context and reasons, and needs restructuring so the answer leads.
• Preparing an executive summary, a decision memo, or the spine of a recommendation deck.

When NOT to Use

• Early exploratory thinking, before a conclusion exists. Answer-first structure forces a premature headline; do the thinking first, then structure the answer.
• To test whether an argument holds. Auditing reasons, co-premises, and objections for soundness is argument analysis (use think-argument-mapping). The pyramid composes a case; it does not check it.
• To decompose a question for analysis. Breaking a problem into MECE sub-questions to investigate is an issue tree (use think-issue-tree), which structures the question for the analyst; the pyramid structures the answer for the reader.
• To make a thin case look authoritative. Confident structure can lend false weight to weak evidence; a tidy pyramid is not a validated argument.

Instructions

When asked to apply the pyramid principle, follow these steps:

1. State the governing thought. Capture, in one sentence, the single conclusion or recommendation the reader should walk away with. If no conclusion exists yet (the thinking is still exploratory), say so and stop - this is the wrong tool.
2. Draft the key arguments. List the small set of reasons (aim for three to five) that, taken together, justify the governing thought. Each must directly answer the question the governing thought provokes (“why do you say that?”).
3. Make the set MECE. Test the key arguments for overlap (mutually exclusive - merge or re-cut any that restate each other) and for gaps (collectively exhaustive - add anything material the set is missing). Keep it small; do not pad.
4. Order the key arguments deliberately. Arrange them by one intelligible logic - importance, time sequence, or structure - not in the order they occurred to you. State which ordering you used.
5. Attach supporting evidence. Under each key argument, list the facts, data, or sub-points that support it. This is where detail lives, so the top stays scannable.
6. Check the logic both ways. Verify vertically (each level answers the one above) and horizontally (the key lines actually sum to the governing thought - no more, no less). Fix the structure if they do not hold.
7. Optionally frame the intro with SCQA. If an opening hook helps, write a Situation the reader accepts, the Complication that disturbs it, the Question it raises, and the Answer (= the governing thought).
8. Emit the pyramid. Produce the artifact in references/TEMPLATE.md: the governing thought, the ordered key lines, and the supporting evidence under each, as an explicit tree.

Output Format

Use the template in references/TEMPLATE.md. The deliverable is the filled pyramid (governing thought, ordered key lines, supporting evidence, optional SCQA intro), not a prose essay.

Quality Checklist

Before finalizing, verify:

• The governing thought is a single, concrete conclusion stated first - not a topic, a question, or a teaser.
• There is a small set of key arguments (roughly three to five), each answering “why?” for the governing thought.
• The key arguments are MECE: no two overlap, and nothing material to the claim is missing.
• The key arguments are in a deliberate order (importance, time, or structure), and the ordering logic is stated.
• The key lines actually sum to the governing thought - they justify it, no more and no less.
• Supporting evidence sits under the right key line, keeping the top level scannable.
• The output is the pyramid artifact, not prose.
• No overclaiming: the skill makes the recommendation clearer to follow; it does not certify that the recommendation is correct (see evidence/dossier.md).

Evidence

Tier P (practitioner). The pyramid principle is a widely and durably adopted convention for executive communication (Minto 1987); its core “state the answer first” move is consistent with general reading-comprehension findings on stating the main point up front. There is no body of controlled studies on the named method, and that comprehension research is adjacent (it tested thesis-first prose in general, not the pyramid, not business recommendations). The evidence is transferred from human practice, not validated for AI-augmented use. The honest value is mechanical: it makes the agent lead with the conclusion and surface an inspectable structure, rather than burying the point. It does not make the recommendation correct. Full grading, sources, and caveats: evidence/dossier.md.

Examples

See references/EXAMPLE.md for a completed pyramid on a real recommendation.

Deep dive: worked example

A full worked run (the shared Northwind scenario)

Pyramid - Worked Example

A completed run of the pyramid-principle skill on a real recommendation. This is the quality bar a generated pyramid should meet.

Use the shared recurring scenario (Northwind, a B2B SaaS weighing a self-serve free-tier launch) so examples across skills read as one coherent product. See docs/internal/AUTHORING.md.

---

Subject

• Recommendation being communicated: Northwind has finished its analysis and decided to launch the self-serve free tier, but only behind explicit guardrails. The exec team now needs that recommendation written up as a decision memo. (The decision is made; this is the write-up, not the analysis.)
• Reader: Northwind’s exec team and the board sponsor for the Q3 growth target - busy readers who want the recommendation first and the option to descend for detail.
• Ordering logic for the key arguments: Importance. The reader’s first question is “should we do this at all?”, so the arguments run growth case, then the conditions that make it safe, then the cost of waiting.

Governing thought (top of the pyramid)

Launch the self-serve free tier in Q3, but only behind three pre-committed guardrails (paid-feature gating, a usage cap with a cost tripwire, and a redesigned sales comp model); without them, hold.

Optional SCQA intro

• Situation: Northwind’s growth is sales-led and predictable, and the Q3 board target assumes a step-change in top-of-funnel volume.
• Complication: The sales-led motion cannot deliver that step-change on its own in one quarter, and competitors now offer a free entry point that is capturing the developers who later choose tooling.
• Question: Should Northwind launch a self-serve free tier this quarter, and if so, on what terms?
• Answer: Yes - launch in Q3, but only behind the three guardrails (= the governing thought).

The pyramid

GOVERNING THOUGHT: Launch the self-serve free tier in Q3, but only behind three pre-committed
                   guardrails; without them, hold.

  KEY ARGUMENT 1: The free tier is the only lever that can hit the Q3 top-of-funnel target.
    - support: Sales-led sign-ups have grown ~12% per quarter; the board target needs ~3x volume, which the current motion cannot reach in one quarter.
    - support: Two direct competitors launched free tiers in the last year and are now first to reach developers who later pick the team's tooling.
    - support: A self-serve path removes the sales-touch bottleneck that currently caps how many small accounts can even enter the funnel.

  KEY ARGUMENT 2: It is only safe if three risks are pre-committed against, so the launch must carry guardrails as conditions, not hopes.
    - support: Cannibalization - without gating the top value features behind paid, existing customers can downgrade to free; guardrail = gate the top three features and instrument the free-to-paid funnel before launch.
    - support: Cost runaway - unqualified free users can swamp support and infra; guardrail = a hard usage cap plus a cost-per-free-user tripwire set before launch.
    - support: Sales conflict - reps will undercut the motion if comp and lead-routing are unchanged; guardrail = a redesigned comp model agreed with sales leadership before any announcement.

  KEY ARGUMENT 3: Waiting a quarter costs more than a guarded launch risks.
    - support: Each quarter of delay cedes the developer-entry point to the two competitors, who compound their lead.
    - support: The guardrails make the downside bounded and reversible-on-signal (caps, tripwires, kill criteria), whereas a missed Q3 target is a fixed, public miss against the board.
    - support: The build cost of the guarded thin-slice launch is small relative to the pipeline it is forecast to open.

Structure check

• Vertical: each key argument answers “why launch in Q3 behind guardrails?” - because it is the only lever that hits the target (1), because it is safe only if guarded (2), and because waiting costs more than guarding (3). Each support backs its own key line.
• Horizontal (MECE): the three key lines do not overlap (the growth case, the safety conditions, and the cost of delay are distinct). Together they cover the reader’s three real questions - is it worth doing, is it safe, and why now - leaving no material gap before a launch decision.
• Sum: the three key arguments justify exactly the governing thought, including its “without them, hold” condition (argument 2 is what makes the guardrails non-negotiable). They do not over-claim a guaranteed outcome.

---

Note how the value is in the inversion: the recommendation and its single hard condition (“only behind three guardrails; without them, hold”) land in the first line, and the reader can stop there or descend exactly as far as their trust requires - whereas the same content told discovery-first would bury the decision under three paragraphs of market context. The pyramid makes the case clearer to follow; it does not, on its own, prove the decision is right.

Grounding: the full evidence dossier

What the research does and does not show, with graded sources

Evidence Dossier: Pyramid Principle

The single source of truth for the pyramid-principle skill. The SKILL.md, the sidecar (skill.meta.yml), and the eval cases all derive from this file. If a claim is not here, it does not belong in the skill.

Skill	thinking-framework-skills.pyramid-principle (installable name think-pyramid-principle)
Family	synthesis
Evidence tier	P (practitioner; limited controlled evidence - see section 3)
Confidence	High that the structure makes a recommendation easier to follow; low that any controlled study has measured it
Status	draft (authored 2026-05-31 from the discovery corpus)

---

1. The mechanism (what actually does the work)

Most recommendations are delivered the way they were discovered: context, then analysis, then (eventually) the conclusion, so the reader has to hold a pile of facts in working memory and wait to learn what they add up to. The pyramid principle inverts that. The single conclusion - the governing thought, the one thing the reader should do or believe - goes first. Beneath it sit a small set of key arguments (typically three to five) that, taken together, justify the governing thought. Beneath each key argument sits the supporting evidence. The result is a top-down tree the reader can descend exactly as far as their trust requires and stop.

Three constraints do the actual work:

1. Answer first. Leading with the conclusion lets the reader evaluate everything below it against a known claim, instead of reverse-engineering the point from the evidence. This is the load-bearing move.
2. A small, MECE set of key arguments. The supports for the governing thought should be Mutually Exclusive (no overlap, so the reader is not re-reading the same point) and Collectively Exhaustive (nothing material to the claim is missing). MECE plus “small” is what keeps the case both complete and followable.
3. Vertical and horizontal logic. Vertically, each level answers the question the level above provokes (“why do you say that?”). Horizontally, the key arguments are ordered by a single intelligible logic (importance, time, or structure), not dumped in the order they occurred to the author.

An optional SCQA opening (Situation the reader accepts, Complication that disturbs it, Question it raises, Answer = the governing thought) gives the introduction a hook that lands on the conclusion rather than wandering toward it.

The mechanism we implement is: state the answer, decompose it into a small MECE set of ordered arguments, attach evidence to each, and pressure-test the structure. The branded “Minto Pyramid Principle” is the packaging; the durable move is answer-first, grouped, ordered communication of a conclusion.

2. Lineage

• The Pyramid Principle: Minto, B. (1987). The Pyramid Principle: Logic in Writing and Thinking. Originated at McKinsey in the 1970s as house guidance for structuring consulting recommendations; it remains the dominant convention for executive communication in management consulting.
• MECE (mutually exclusive, collectively exhaustive): a grouping discipline popularized alongside the pyramid in the same consulting lineage.
• SCQA (situation-complication-question-answer): Minto’s introduction pattern, widely taught as a standalone framing device.

“Pyramid Principle” is a book title and a widely used generic phrase for the technique; “Minto” is a personal name and an informal label for the method. We name the skill descriptively (pyramid-principle) and cite Minto in the lineage rather than branding the skill. No attribution is required for the generic technique.

3. What the evidence shows, and what it does NOT show

This is the honest core of the dossier. The skill must not overclaim.

What is reasonably supported (the practitioner case):

• The method is widely and durably adopted in management consulting, corporate strategy, and executive communication - decades of practitioner use across firms, which is real signal that it is useful for its job (communicating a recommendation to a busy reader).
• Its core move - state the conclusion first - is consistent with well-established reading-comprehension findings that an explicit topic/thesis stated up front (an “advance organizer”) helps readers comprehend and recall structured expository text. The broad direction (“tell readers the point first”) is supported by comprehension research even though that research did not test the pyramid method itself.

What is NOT shown (the caveats that keep the skill honest):

• There is no body of controlled studies on the Pyramid Principle as a named method. It is practitioner doctrine, not an experimentally validated intervention. We grade it P, not S or M, for exactly this reason.
• The comprehension research that supports “answer first” is adjacent, not direct: it studied advance organizers and thesis-first prose in general, not Minto’s pyramid, not business recommendations, and not AI-generated ones. Treat it as a plausibility anchor, not proof.
• The method improves communication of a conclusion the author already holds. It does not test whether the conclusion is correct, and it can make a weak argument sound more authoritative by dressing it in confident structure. A clean pyramid is not a sound argument; that is a job for argument analysis, not for this skill.

Net grade: P. Useful, heavily field-tested practitioner method with a comprehension-research direction behind its central move, but no controlled validation of the method as such. Claim the communication benefit; disclaim any claim that it makes the recommendation correct.

4. Transferred-evidence flag (required honesty for this library)

All of the support above comes from human practice and human-subject comprehension research, none of it from AI-augmented use. There is no study of a pyramid built by, or with, an AI agent, nor of whether an agent-produced pyramid improves a human reader’s decision. The evidence is therefore transferred from human contexts, not validated for AI-augmented use. This skill must say so.

The realistic AI value does not depend on the unproven claims: a model defaults to narrating its reasoning (context first, conclusion buried, supports in discovery order). This skill makes the agent invert that default - lead with the governing thought, force the supports into a small MECE set, order them deliberately, attach evidence, and surface the structure as an inspectable artifact a human can challenge. That is a reliable, mechanical improvement to how a recommendation is communicated, independent of whether anyone has measured the method in a lab.

5. When it works / when it fails (drives the eval negative cases and “When NOT to Use”)

Works best when:

• There is already a conclusion or recommendation to communicate (the thinking is done; this is the write-up).
• The reader is busy or senior and needs the headline first, with the option to descend for detail.
• The supports can be grouped into a small, non-overlapping set and put in a deliberate order.

Fails or misleads when (poor-fit / anti-patterns):

• There is no conclusion yet. Early exploratory thinking, where the answer is genuinely unknown, is the wrong place for an answer-first structure - it forces a premature headline. (Anti-trigger.)
• The task is to test whether an argument holds, not to communicate one. Auditing reasons, co-premises, and objections for soundness is argument analysis (use think-argument-mapping); the pyramid composes a clear case, it does not check it. (Near-miss anti-trigger.)
• The task is to decompose a question for analysis, breaking a problem into MECE sub-questions to investigate. That is an issue tree (use think-issue-tree); it structures the question for the analyst, whereas the pyramid structures the answer for the reader. The two look similar (both are MECE trees) and are easily confused. (Near-miss anti-trigger.)
• Run as ritual - slapping a one-line headline on top of unchanged, ungrouped prose, with supports that overlap or leave gaps, and key lines that do not actually sum to the governing thought. A pyramid whose levels do not hold together is cargo-cult structure. The skill must enforce MECE and the “do the key lines justify the top?” check.
• Used to make a thin case look strong. Confident structure can lend false authority to weak evidence. The skill must not imply that a tidy pyramid is a validated argument.

6. Output artifact

The skill must emit a pyramid, not prose: a governing thought at the top, a small ordered set of key-argument lines beneath it, and the supporting evidence under each key line, plus (optionally) an SCQA intro framing. Represent it as an explicit outline/tree (indented levels or a small table), preceded by a one-line statement of the governing thought so a reader who stops at the top still has the recommendation. The structure is the deliverable; a flowing essay is not.

7. Sources

1. Minto, B. (1987). The Pyramid Principle: Logic in Writing and Thinking. - the method, MECE grouping, and SCQA introduction; originated as McKinsey house guidance.
2. Reading-comprehension / advance-organizer literature (e.g. Ausubel’s advance-organizer work and thesis-first expository-text studies) - the general finding that stating the main point up front aids comprehension and recall of structured text. Cited as an adjacent plausibility anchor for the “answer first” move, not as a test of the pyramid method.

Verification status: Minto (citation 1) is the well-attested primary source for the method and its components. The comprehension/advance-organizer link (citation 2) is drawn from secondary synthesis and is deliberately framed as adjacent support, not direct validation; confirm the specific studies against primary sources before any public-facing claim, and never upgrade this from a plausibility anchor to evidence that the pyramid method itself was tested. The “no controlled studies of the named method” statement in section 3 is the honest default and should stand unless a primary study is found.