Full SKILL.md
Executive Communication
Turns finished or near-finished analysis into a client-ready argument using Minto's pyramid and SCQ — governing thought, key line with checked horizontal logic, dot-dash storyline, action titles that pass the tests, ghost deck, one-page executive summary or memo, objection table, speaking notes and pre-wiring plan. Use whenever a CEO, board, steering committee or investor must read or hear analysis and decide, or the user says "build the storyline", "write the exec summary", "what's the headline", "turn this into a deck", "fix my action titles", "ghost deck", "it doesn't flow", "prep me for the steerco", "how do I land this with the CEO", "what will they push back on". Produces content and structure only; pptx and docx render the file. Not for choosing what to analyse (problem-framing).
Produces the argument, titles, slide-by-slide content plan and summary text; pptx and docx render them. Minto's grouping rules, deductive vs inductive support, and Aristotle's three appeals are in references/frameworks.md; read it only when a step needs defending.
When not to use
- Analysis not done and the question is what to analyse:
problem-framing,hypothesis-stress-testing. - Research to support the argument:
boardroom-storm-research. - The file itself:
pptx,docx. A product-strategy deck:product-strategyholds the deck architecture and storyline blocks; call this skill for the pyramid, titles and objection prep.
Inputs
Ask for, or infer and label: the decision in one sentence and who makes it; the audience (names, roles, current beliefs, sceptics, who is briefed, time allotted, read alone or presented); the analysis (findings, numbers, sources, confidence, what was not analysed); the original SCQ if one exists; format (deck, memo, one-pager, spoken, combination); constraints (house style, page limits, sensitivities, public prior commitments).
Ask one round of questions if the decision or the analysis is missing. Otherwise assume a CEO plus two or three directs, a thirty-minute slot, a deck read in advance and presented; put assumptions in a short block after the governing thought.
Workflow
-
Question and audience (
references/templates.md§1). The question in the audience's words, not the team's. Audience table, at most eight rows: person, role in the decision (decides, must agree, advises, executes), current belief, what they lose or gain, evidence they trust, briefed or not. Write for the decider; handle the others in pre-wiring. Test: if the CEO reads only the first sentence, is it the answer to the question they asked? -
SCQ. Situation the reader already accepts (no new information); Complication that makes it unstable; the Question it raises. If it differs from Stage 1, reconcile. Test: would the reader agree with every word before seeing any analysis?
-
Governing thought. One sentence under 35 words with a verb of recommendation and a number, that could be false ("recover margin through pricing and mix rather than cost, worth $22–31M in 24 months"). Elevator test: with one reason, in thirty seconds, would a CEO know what is being asked?
-
Key line and horizontal logic (
references/templates.md§3). Two to four ideas supporting the governing thought. Deductive (A, B, therefore C; compelling, fragile; at most four steps, never nested at one level) when the audience must follow a chain; inductive (same-kind ideas under a plural noun; survives losing a leg) when reasons are independent. Per group: name the relationship; if deductive confirm B comments on A and C follows; if inductive write the plural noun, confirm every member fits and none is missing, state the order (time, structure, degree); the summary above is a claim the members prove, not a count ("there are three reasons" fails); the members answer exactly the reader's next question (why, how, how much) and nothing else. Most strategy storylines are deductive at the top, inductive beneath. Test: remove any key-line point; does the argument stand or collapse, and is that intended? -
Dot-dash storyline (
references/templates.md§4). Each dot one slide's message as a full sentence (the future action title); under it two to four dashes of evidence with the chart that shows it. Twelve to fifteen dots for a thirty-minute decision meeting. A dot with no dashes is unprovable: find evidence or cut. A dash no dot claims goes to appendix or nowhere. Mark each dash have / pending / assumed; pending dashes are the workplan. Test: read the dots alone; complete argument, answer first? -
Action titles (
references/templates.md§6). Convert each dot; rewrite any that fails: a claim not a label; one idea (cover the body, the title stands; cover the title, the body proves only that); carries the so-what; specific enough that it could not sit over another client's slide; 25 words or fewer, counted (write counts in working notes, delete before delivery); every number provable on the slide; reads in sequence without "this" or "the above"; no "overview of", "key insights", "deep dive". Titles must add up: per-lever titles may not sum to more than the headline; no title claims a gap closes fully unless the analysis says so; unsized levers carry direction and a marked assumption, not a figure. -
Ghost deck (
references/templates.md§5). Per slide: title, exhibit sketch (chart type, axes, the one comparison the eye must make), source, status. Every analysis feeds a slide or is cut. Appendix list. Test: which slides are still ghosts, and what happens to the argument if their data comes back the other way? -
Executive summary and memo (
references/templates.md§7–8). One page: SCQ in a paragraph; governing thought as the first sentence of the second paragraph; key-line points numbered with their strongest number; decisions requested with dates; risks and unknowns. Readable in ninety seconds; matches deck titles word for word where they overlap; introduces no number not in the deck; hedges exactly as much as the deck. Memo: same structure, one to two pages, more support per point. -
Ethos, pathos, arrangement. Ethos: method briefly (sample sizes, reconciliation to audited figures, exclusions), honest confidence, what was not done; describe only a data basis the user gave, otherwise "data basis: to confirm with [owner]" on the gaps list. Pathos: the cost of doing nothing, in the audience's terms, once. Refutation: where the strongest counter-argument is acknowledged and answered. Annotate the storyline with where each lands.
-
Objections, speaking notes, Q&A (
references/templates.md§9–10). Objection table, six to ten rows: decision-maker, objection, answering evidence, slide, fallback if unpersuaded, and what must not be offered. Each objection resolves one of three ways: answered on a slide; a real risk moved to the risks slide with mitigation and owner; a gap that sends the argument back to Stage 4. Speaking notes per main-body slide. Test: hardest question, who asks it, which slide? -
Pre-wire (
references/templates.md§11). One-to-ones before the meeting: allies and data owners first (confirm numbers), sceptics next (objections heard in private, absorbed where valid), most senior last, briefed on what the others said. Exception: when the recommendation reverses a decision someone made, a short no-document heads-up to them first. Plan: who, when, what to show, expected objection, outcome. Revise the storyline afterwards. -
Deliver and hand off. Deliverable first; working notes after, only if asked or if they change the argument. Titles unchanged through rendering; if a designer shortens one, re-run the Stage 6 tests.
Output
The governing thought is the first sentence. Then the short assumptions block if any, then by request type:
- Storyline: 12–15 dots with dashes and status; executive summary under 400 words; objection table; pre-wiring plan; speaking notes only if asked. 1,500 to 3,300 words with tables.
- Titles only: rewritten titles; one paragraph on what changed and why; per-slide notes only where proof is not obvious. Under 1,100 words.
- Memo: the memo, then objection table.
Four items always included and protected from trimming: the thirty-second version (governing thought plus one reason); a "Gaps to close before the meeting" list, at most six lines with owners; the fallback per hard objection; the pre-wiring order. No method vocabulary, framework authors, skill names or the worked example's client in the deliverable. Full example: references/worked-example.md.
Quality bar
Run silently; fix, do not report.
- Governing thought is first, under 35 words, has a verb and a number, and could be false.
- Situation and complication contain nothing the reader disputes.
- Every group is named deductive or inductive and passes its check; no blank summaries.
- Dots alone read as a complete argument; no dot without dashes.
- Every title passes all eight tests and is 25 words or fewer, counted.
- Titles sum correctly to the headline; no unsupported closure claims.
- Summary introduces no new numbers and hedges as the deck does.
- Every objection resolves to slide, risk or gap.
- Data basis is stated only where the user supplied it.
- The thirty-second version, gaps list, fallbacks and pre-wiring order are present.
- No leaked method vocabulary; no repeated titles.
Anti-patterns
- Answer on page 30; opening with assumptions or an audience table.
- Topic-label titles; "there are three reasons" summaries.
- Deductive all the way down (a proof that loses the room at step two); inductive all the way down (a list with no recommendation).
- Titles that cite figures only in the appendix; per-lever titles that outsum the headline.
- A summary that hedges differently from the deck.
- An invented method line.
- Nothing pre-wired; the sponsor surprised in the room.
Hand-offs
References
Frameworks: the intellectual foundation of executive communication
4,359 wreferences/frameworks.md6 sub-sectionsTemplates and checklists for executive communication
1,680 wreferences/templates.md20 sub-sectionsWorked example: Northgate Industrial Supply margin recovery
4,048 wreferences/worked-example.md10 sub-sections