From grimoire
Structures writing with the conclusion first, followed by grouped supporting arguments. Speeds comprehension in business documents, memos, emails, and presentations.
How this skill is triggered — by the user, by Claude, or both
Slash command
/grimoire:apply-pyramid-principleThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Lead with the answer, then support it with grouped, ordered reasoning — never make the reader wait for the conclusion.
Lead with the answer, then support it with grouped, ordered reasoning — never make the reader wait for the conclusion.
Adopted by: McKinsey & Company (where Minto developed it — required training for all associates and consultants), BCG, Bain, Goldman Sachs, Morgan Stanley, Deloitte. Harvard Business School includes it in core curriculum. Standard for all client-facing deliverables in top-tier consulting and investment banking globally for 50+ years. Impact: Top-down structure cuts time-to-comprehension because readers get the conclusion before processing supporting detail — the brain can file each argument under a known conclusion rather than holding uncertainty in working memory. Bottom-up writing (conclusion last) forces readers to re-read once the conclusion arrives. McKinsey measures communication quality on this framework in internal reviews. Why best: Most writers default to bottom-up (chronological story → conclusion) because that's how thinking unfolds. But readers are not interested in the writer's thinking process — they want the answer. Pyramid Principle forces the discipline of separating thinking order from communication order.
Sources: Barbara Minto, "The Minto Pyramid Principle" (1987); McKinsey & Company internal writing standards; Harvard Business School MBA core curriculum
[Governing Thought — the single answer]
/ | \
[Key Point 1] [Key Point 2] [Key Point 3]
/ \ | / \
[Data] [Data] [Data] [Data] [Data]
Every level answers the question raised by the level above it. The governing thought answers: "So what should I do/know/think?" Key points answer: "Why?" or "How?" Supporting data answers: "How do you know?"
Write the single sentence that is your entire message. If forced to telegram the document, this is the telegram.
Bad: "There are several factors affecting our Q3 performance..."
Good: "We should exit the German market by Q4 to stop the cash drain."
This goes at the top — title, subject line, or opening sentence. Never at the end.
When the reader doesn't yet accept the problem framing, use SCQA before the governing thought:
Situation: Our German operations have grown to €12M revenue over 5 years.
Complication: The market requires 3× our current investment to reach profitability,
and two local competitors have locked up the top three accounts.
Question: Should we continue investing or redirect resources?
Answer: We should exit Germany by Q4 and redeploy capital to the UK expansion.
Skip SCQA when the reader already accepts the problem — go straight to the answer.
Supporting points must be:
Bad grouping (overlapping):
- Cost is too high
- Margins are too low ← same problem, two framings
- Revenue doesn't cover costs ← same again
Good grouping (MECE):
- Unit economics are structurally negative (cost > price ceiling)
- Market share is unwinnable (top accounts locked by incumbents)
- Capital requirement exceeds strategic value (3× invest for 15% share)
Aim for 3–5 key points. Fewer than 3 feels thin; more than 5 overloads working memory.
Choose one ordering principle and apply it consistently within each level:
| Order type | When to use | Example |
|---|---|---|
| Deductive | One argument flows from the previous | Major premise → minor premise → conclusion |
| Chronological | Steps or phases | Step 1 → Step 2 → Step 3 |
| Structural | Physical or org components | North region, South region, West region |
| Importance | Independent parallel points | Most critical → least critical |
Don't mix ordering types within the same group.
Read each level and ask: "So what does this mean for the level above?" If the answer isn't already in the document, the logic is incomplete.
Data: "Customer churn increased 8% in Q3."
So what? "Our retention problem is accelerating." ← this must appear as a key point
So what? "We must fix retention before expanding." ← this must appear at governing thought
| Format | Pyramid location |
|---|---|
| Subject line = governing thought; first paragraph = key points | |
| Memo | Title + opening paragraph = governing thought + SCQA; sections = key points |
| Slide deck | Title slide = governing thought; each slide title = key point (not topic) |
| Verbal briefing | First sentence = governing thought; then key points with evidence |
Slide titles must be assertions, not topics:
Bad title: "Market Analysis"
Good title: "German market cannot reach profitability within our investment horizon"
Bottom-up structure. "We analyzed X, then Y, then Z, and therefore we recommend…" The reader waited through the entire analysis to get the answer. Invert it.
Topic slide titles. "Revenue Trends" tells the reader nothing. "Revenue growth is decelerating and will miss target by 15%" tells them everything before they read the chart.
Non-MECE groupings. Two key points that overlap confuse readers and signal fuzzy thinking. If points overlap, merge them or find the higher-level point they both support.
Governing thought too vague. "We need to improve our go-to-market approach" is not a governing thought — it's a topic. "We should hire three enterprise sales reps in H1 to capture the mid-market segment before the competitor launches" is a governing thought.
npx claudepluginhub jeffreytse/grimoire --plugin grimoireCreates polished client-facing consulting documents and presentations following top-tier consulting conventions. Applies pyramid principle, visual hierarchy, quantified impact, and content discipline to proposals, reports, and board decks.
Creates consulting-grade reports and executive presentations using SCQA/SCR structure, Pyramid Principle for strategic assessments, board decks, and due diligence.
Plans persuasive presentation structure: audience analysis, narrative arc selection (SCQA/problem-solution), and evidence sequencing for talks and pitches.