Source: frameworks/kit-li-thought-leadership/04-li-thought-leadership-quality.md

LinkedIn Thought Leadership Post — Quality Control

Pre-Build: Status Viability Gate

Before anything else — before intake, before writing — confirm this topic can produce a status-positive LinkedIn post:

[ ] Can complete: "The reader who engages is telling their network: ___"
[ ] That signal is something the reader WANTS their network to see
[ ] Engaging signals capability, sophistication, or leadership — NOT a problem, weakness, or gap
[ ] A practice owner making $500K–$2M would feel good about their peers seeing their comment

If ANY of these fail, stop. Pivot the angle, move to a different format (article, Substack, HR post), or pick a different concept. Do not proceed to intake.

Pre-Build: Intake Validation

Before writing starts, confirm a Post Lock exists with all seven decisions made:

[ ] Angle locked (from the angle library)
[ ] Core insight articulated — one sentence, names what's counterintuitive
[ ] Status angle articulated — what does engaging signal about the reader?
[ ] Hook locked — stops the scroll, works above the fold
[ ] Arc defined — section flow matches the angle
[ ] Closer locked — quotable, sharp, statement not question
[ ] Conversation prompt locked — specific, status-positive for commenter
[ ] Format locked — text post or carousel

If any of these are missing, run the intake before writing.

Image Quality

[ ] Image passes the 2-second test — someone with zero context understands it in 2 seconds
[ ] Image stops the scroll — it does NOT deliver the concept or summarize the post
[ ] Uses the approved quote card format: primary text (ivory) + gold divider + secondary text (gold)
[ ] Primary text is 108px+ at 1200px canvas (reads on a phone in the LinkedIn feed)
[ ] Visual brand is consistent with the campaign (canvas, background, fonts, colors, logo)
[ ] Campaign label is present (top center, small tracked uppercase)
[ ] If the concept requires explanation → quote card, not data visual
[ ] No crossing-lines visualizations, data tables, numbers without context, or novel frameworks as visuals

Hook Quality

[ ] Hook is visible before "See more" fold (~210 characters max for first 2 lines)
[ ] Hook creates behavioral recognition — reader thinks "that's me," not "tell me more"
[ ] Hook names a behavior the reader performs, not a conclusion they should draw
[ ] Hook uses one of the five types: Scene, Claim, Number, Quote, or Pattern
[ ] Hook does not start with "I" (first-person openers underperform on LinkedIn)
[ ] Hook is specific, not vague ("A wealth advisor lost her second-largest client" not "Losing clients is painful")
[ ] No question hooks — questions give the reader permission to answer "no" and scroll past
[ ] Hook signals status, identity, or stakes — clicking "See more" is already a self-selection act
[ ] Hook is one line, under 15 words
[ ] Hook uses the reader's internal vocabulary, not consultant language ("that quarterly report" not "service delivery")
[ ] Hook lands without setup — no second line needed to make it work
[ ] Hook does NOT: give a conclusion, give away the secret, make a prediction, use dramatic reveal/ellipsis setup

First Line Quality

[ ] First line sets up proof — opens a list, evidence, or framework
[ ] First line is under 10 words
[ ] No transition language ("Let me show you...") — just opens what comes next
[ ] No elaboration on the hook — the hook already landed
[ ] No second hook — don't re-hook when the first hook worked
[ ] Content starts immediately after the first line — no blank line drama

Distribution Architecture

Dwell Time

[ ] Total length is 1,300–1,900 characters (250–400 words) — the algorithm sweet spot
[ ] Post is not under 500 characters (algorithm flags as low-effort)
[ ] Post is not over 3,000 characters (completion rate drops)
[ ] At least one section has escalating specifics or concrete details that slow the reader down
[ ] The body rewards reading — each block reveals something the previous one set up
[ ] If carousel: 6–12 slides, one idea per slide, reader must swipe to get the full argument

Comment Generation

[ ] Conversation prompt exists at the end of the post
[ ] Prompt invites a specific, substantive response (not "Agree?" or "What do you think?")
[ ] Prompt is status-positive for the commenter — answering makes them look smart, experienced, or thoughtful
[ ] Prompt requires multi-sentence answers (not yes/no)
[ ] Prompt has thread potential — one person's answer could trigger another person's response
[ ] Prompt uses one of the six types: Experience mirror, Forced choice, Debate invitation, Application challenge, Turning point, or Specific ask

Algorithmic Hygiene

[ ] No external links in the post body (60% reach penalty)
[ ] No "link in comments" references
[ ] No hashtags (Advisory OS doesn't use them — they dilute authority positioning)
[ ] No emojis
[ ] No engagement bait phrasing ("Tag someone," "Share if you agree," "Comment YES")
[ ] Post stays in the Advisory OS topic lane (operational systems for advisory practices)
[ ] Content includes genuine human insight — not AI-reorganized information

Status and Identity Check

[ ] The status angle is clear — you can articulate what engaging signals about the reader
[ ] Reader finishes the post feeling elevated, not lectured
[ ] The insight gives the reader language or a framework they can use with others
[ ] Engaging (liking, commenting, sharing) is a status-positive act for the reader
[ ] No pandering ("If you're reading this, you're ahead of 90%")
[ ] No status-negative conversation prompts (don't ask people to admit failure publicly)
[ ] The post gives the reader something they can repeat in their own conversations — that's the status transfer

Structural Quality

[ ] Post follows the locked angle's default arc (or a justified variation)
[ ] Post follows Steve's DNA: short lines, white space, scannable on mobile
[ ] Each number, each beat, each statement gets its own line
[ ] No narrative essays or paragraphs of prose
[ ] Each block (1–3 sentences) does one thing
[ ] The turn or pivot moment is clear
[ ] Evidence is specific: names, numbers, dialogue, or concrete details
[ ] Single through-line — one insight, not three
[ ] No filler blocks — removing any block would weaken the argument
[ ] Closer lands before the conversation prompt
[ ] Closer is quotable without context
[ ] One scenario per post — no character switching, no second example introduced
[ ] Every pronoun clearly tied to one established character — no ambiguous he/she
[ ] Establish role AND action before switching to pronouns
[ ] Numbers consistent throughout — no unexplained shifts between figures
[ ] If multiple examples needed, the article handles it — the post gets one

Prose Quality

Voice Compliance (voice.md)

[ ] Vocabulary uses: deploy, build, install, constraint, bottleneck, infrastructure, capability, system
[ ] Vocabulary avoids: leverage, synergy, scalable, coaching, accountability, mindset, crush it, game-changer, level up, journey, transformation, revolutionary, groundbreaking
[ ] Tone is confident and direct — surgeon, not guru
[ ] Paragraphs stay tight — 1–3 sentences per block
[ ] Specific numbers over vague claims
[ ] No hedge language where the pattern is clear
[ ] No academic phrasing, no marketing hype, no self-deprecating qualifiers

Sentence-Level (linkedin-sentence-editor.md)

[ ] Rule 1: End strong — last word is the most impactful
[ ] Rule 2: Acronyms defined on first use
[ ] Rule 3: Economy — no wasted words
[ ] Rule 4: No repetition — every sentence says something new
[ ] Rule 5: Key words varied
[ ] Rule 6: 3–5 precision vocabulary upgrades
[ ] Rule 7: Two-comma max per sentence
[ ] Rule 8: Adverbs killed

AI Pattern Detection (copy-qc.md)

QC posture: fail-first. When a pattern is detected, the default is FAIL. To keep a flagged line, cite a specific golden example (03-golden-examples.md) that uses the same structure and was annotated as acceptable in its Copy QC Notes. No rationalizing — show the precedent or rewrite.

P1 — Must Fix

[ ] No twinning ("You don't have X. You have Y.")
[ ] No mirror reversals
[ ] No question → dramatic silence → revelation arcs
[ ] No correction-revelation family compounds (2+ from Patterns 1/3/4/7)

P2 — Rewrite Unless Exception

[ ] No three-beat parallel lists (unless data pattern)
[ ] No "Not because X. Because Y."
[ ] No formulaic setups — max 1 per post with real evidence
[ ] No trying-to-be-quotable lines (metaphor over clarity)

P3 — Fix If Multiple

[ ] Max one dramatic single-word beat
[ ] No rhetorical hand-holding ("Right?" "Sound familiar?" "Let that sink in.")
[ ] No three+ consecutive same-opener sentences (unless data pattern)

Compound Check

[ ] No pattern accumulation creating overall AI feel
[ ] Sentence lengths vary
[ ] Passes read-aloud test

LinkedIn Platform Quality

[ ] Line breaks used for rhythm, not as default
[ ] No more than 3 consecutive single-sentence lines without a multi-sentence block
[ ] Dialogue lines each on their own line
[ ] Post works if someone only reads the hook and closer
[ ] If carousel: slides are 1080×1350px, charcoal/cream/gold palette, one idea per slide

Structural Gate

[ ] Post is scene-led, not thesis-led — the insight emerges through narrative, not announced in the hook
[ ] Structural pattern matches the closest golden example (03-golden-examples.md)
[ ] If thesis-led: rewrite scene-first before running any other checks

Final Gate

[ ] Would the writer say every line of this in a real conversation?
[ ] Does this post name a pattern, or just describe a symptom?
[ ] Could a competitor copy-paste this? (If yes, it's not distinctive enough)
[ ] Can you see it in your head? (Concrete beats abstract)
[ ] Does the closer land harder than the hook?
[ ] Would a smart practice owner feel good about their network seeing their comment on this post?
[ ] Will this post generate multi-sentence comments in the first 60 minutes?

Review Notes

v1 — First pass from Post Lock. Should be 90%+.
Ship — Passes all checks. Copy QC clean. Sentence Editor clean.