Who you're writing for and how they reason — reader profile, motivation, reasoning, receptivity, response, plus platform and genre calibration. From the Audience Playbook.
Identify the reader's professional or community role and their level of expertise in the post's domain.
Outputs
Role profile with expertise calibration and assumed-knowledge map.
How it differs
Demographics flatten readers; role and expertise differentiate them. A senior researcher and a curious generalist can both read your post and need different things.
Best for
Calibrating how much to explain, what jargon is allowed, what citations belong inline vs in footnotes.
Failure mode
Writing for the imagined "smart generalist" when the actual readership is ~70% specialists who find the explanations patronizing — or vice versa.
Read the platform as a context with norms, recent discourse, and conversational temperature.
Outputs
Platform context map — what's being said, what's tolerated, what triggers what.
How it differs
Platforms are not neutral pipes. The same post on LinkedIn vs Twitter vs Medium will land in different conversational fields. Knowing the field is part of knowing the audience.
Best for
Posts entering active discourse; commentary on recent events; arguments that risk pattern-matching to a current debate.
Failure mode
Posting into a context the writer is not aware of and triggering reactions that have nothing to do with the post's content.
Surface the mental models the reader is likely to bring to the topic — the implicit theories that shape what they'll find plausible.
Outputs
Inventory of likely mental models with compatibility map.
How it differs
Most arguments fail not because the reader rejects a claim but because the reader is using a different model of the underlying domain. Naming the model is the first move.
Best for
Counterintuitive arguments, cross-domain claims, posts that depend on the reader updating a default frame.
Failure mode
Arguing against the wrong model — refuting what you think the reader believes rather than what they actually believe.
Characterize how the reader evaluates evidence and reaches conclusions.
Outputs
Epistemic profile of the audience with calibration recommendations.
How it differs
Readers vary not just in what they believe but in how they form belief. A skeptic-by-default needs different argumentative architecture than a charity-by-default reader.
Best for
Calibrating evidence density, hedge structure, and the where-to-place-the-strongest-evidence question.
Failure mode
Writing as if all readers share the writer's epistemic style.
Estimate how comfortable the reader is with unresolved tension, multiple frames, or honest uncertainty.
Outputs
Ambiguity tolerance assessment with closure-or-openness recommendation.
How it differs
Some audiences want clean answers and read hedging as weakness. Others want honest uncertainty and read closure as overconfident. Same prose, opposite reactions.
Best for
Deciding whether to close the argument or leave it open; how much to qualify.
Failure mode
Forced closure on genuinely uncertain claims (loses the careful reader); or excessive qualification on confident claims (loses the practical reader).
Identify what types of evidence this audience treats as authoritative.
Outputs
Evidence-type ranking for this audience with placement recommendations.
How it differs
"Evidence" is audience-relative. A peer-reviewed study lands with one audience; a personal story lands with another; a working code example lands with a third. The strongest argument uses the evidence the audience actually trusts.
Best for
Deciding which evidence to lead with; which to elaborate; which to relegate to a footnote.
Failure mode
Using the evidence type the writer trusts (typically: studies and citations) regardless of what would convince the reader.
Assess perceived risks and design trust signals matched to specific reader fears.
Outputs
Risk-lever map with positioned trust signals and priority.
How it differs
Generic trust signals (more citations!) miss the specific fears. A reader worried about manipulation needs different signals than one worried about wasted time.
Best for
Posts making strong claims that readers might distrust.
Failure mode
Generic trust signals — "as research shows" — that confirm rather than counter the distrust.
Detect where the post will trigger defensive reactions because it threatens the reader's identity, group, or self-concept.
Outputs
Threat-point inventory with mitigation strategies.
How it differs
Identity threat closes minds against true claims. Detecting it lets the writer choose between (a) defusing it, (b) accepting it as cost, or (c) restructuring to avoid it.
Best for
Posts critiquing practices, communities, or worldviews the reader identifies with.
Failure mode
Triggering identity threat unintentionally and losing readers who would otherwise have agreed.
Predict what the reader will actually do after reading.
Outputs
Action prediction with conversion path analysis.
How it differs
Most posts implicitly assume the reader will do something — share, save, change a behavior, follow up. Naming the assumed action surfaces whether the post is actually designed for it.
Best for
Concluding sections; calls to action; deciding whether the post needs a closing move at all.
Failure mode
Vague hopes ("readers will think about this") substituting for specific predicted actions.
Identify how the post is most likely to be misinterpreted, especially in adversarial or low-attention reception.
Outputs
Misread inventory with preemption strategies.
How it differs
A post does not just have a meaning — it has the meanings other people will attribute to it. Misread risk is about closing the gap between intended meaning and likely uptake.
Best for
Posts that could be screenshot-quoted out of context; nuanced arguments at risk of being flattened; commentary in active debates.
Failure mode
Assuming charitable reading by every reader; failing to consider the worst-faith but plausible interpretation.
- Length: 200-400 words for posts; up to 1500 for articles. Posts longer than ~200 words require strong hook.
- Tone: Professional, often performatively warm. Authentic candor is rare and stands out.
- Structure: Front-loaded hook (first line is the only line guaranteed to be read in feed). Short paragraphs. Whitespace.
- Conventions: Personal stories with workplace lessons. Lists. "I learned this from my mistake" arc. Engagement bait.
- Taboos: Direct political commentary; criticism of named companies; cynicism without redemption arc.
- Reader state: Often phone-bound, between meetings, mid-scroll. High distraction, low willingness to follow complexity.
- Length: 800-2000 words typical; long-form (2000+) accepted with strong hook.
- Tone: Reflective, narrative-tolerant, willing to follow the writer's voice.
- Structure: Headline-driven. Subheadings preferred. Pull quotes. Image breaks for long posts.
- Conventions: Narrative-essay form. "How I came to think X." Tech-meets-humanism. Frameworks named in title.
- Taboos: Bare-bones argument without narrative; pure listicle without redemption; SEO-bait without substance.
- Reader state: Often desktop, deliberate browsing or arriving from share. Medium attention budget; willing to invest if the hook lands.
- Length: Single tweet (~280 chars) or thread (5-15 tweets). Long-form quote tweets travel poorly.
- Tone: Punchy, often confrontational. Voice carries. In-jokes and references signal in-group.
- Structure: Hook tweet must stand alone. Each thread tweet should be quotable individually. Final tweet often has the synthesis.
- Conventions: "Hot take" framing. Numbered threads. Engagement-bait first lines.
- Taboos: Hedging; long throat-clearing; multi-clause sentences in the hook.
- Reader state: Lowest attention budget. Reading in 5-second windows. Willing to engage with one strong claim.
- Length: 1500-4000 words common; some publications go shorter (800) or longer (6000+).
- Tone: Personal, voice-driven, often opinionated. Writer-reader relationship is central.
- Structure: Often opens with personal hook, develops argument, closes with reflection or call. Section breaks common.
- Conventions: Recurring framings ("In this issue..."), reader-direct address, callbacks to previous posts.
- Taboos: Sound corporate; lack of voice; covering ground that could've been a tweet.
- Reader state: Subscribed, opted-in. Higher trust, higher attention, longer commitment than feed-platform readers.
- Length: Variable by subreddit. Technical subs reward long-form (1000-3000 words). Discussion subs reward concise (200-500).
- Tone: Skeptical by default. Specialist subs are merciless to surface-level claims. Hostile to self-promotion.
- Structure: Title is critical. TL;DR often expected for long posts. Sources upfront for contested claims.
- Conventions: Disclosure of credentials/affiliation; explicit acknowledgment of subreddit-specific norms.
- Taboos: Cross-posting without context; appearance of self-promotion; vague claims; unsourced strong assertions.
- Reader state: Adversarial scrutiny by default. Will fact-check. Will downvote weak reasoning visibly.
- Length: Self-determined. Whatever the argument requires.
- Tone: Whatever the writer chooses. No platform pressure.
- Structure: Free. Can use bespoke layout, footnotes, sidenotes, interactive elements.
- Conventions: Linkable. Often referenced from elsewhere — must work for cold-arrival readers.
- Taboos: Few. The main risk is excessive freedom producing structureless drift.
- Reader state: Either deliberate (followed link from somewhere they trust) or zero context (search arrival). Both must be served.
- Reader expectation: Take a position; provide angle the dominant discourse is missing; be timely.
- Reader stance: Already informed about the event; reading for your angle, not the news itself.
- Risks: Hot-take mode (claim without grounding), lukewarm-take mode (rehash without angle), late-take mode (the moment has passed).
- Strong move: Identify the move the dominant discourse hasn't made and make it.
- Reader expectation: Help me understand something I don't currently understand. Build the model in me.
- Reader stance: Knows they don't know. Patient with explanation if it earns the patience.
- Risks: Patronizing the informed reader; losing the novice; using vocabulary without definition.
- Strong move: Worked example or analogy that builds the model in steps.
- Reader expectation: Identify what's wrong with X (a claim, a paper, a movement, a practice). Be sharp, be fair.
- Reader stance: Either an in-group member curious about the critique, or someone the critique might offend.
- Risks: Strawmanning the target; punching down without acknowledging it; cynicism without alternative.
- Strong move: Steelman the target before refuting; offer the alternative position concretely.
- Reader expectation: Take me through your experience and bring me to a generalization that earns its weight.
- Reader stance: Curious about the writer; willing to follow if voice is strong; impatient with self-indulgence.
- Risks: All experience without insight; all insight forced from thin experience; oversharing without point.
- Strong move: Specific lived detail that supports a claim the reader can use.
- Reader expectation: Tell me what the paper says, what's new, what to make of it, whether to read the original.
- Reader stance: Time-constrained, evaluating relevance to their own work.
- Risks: Restating the abstract without synthesis; hagiography of the paper; hidden critiques where a clear judgment would serve.
- Strong move: Sharp claim about what the paper does or fails to do; specific numbers; explicit relevance to a named question.
- Reader expectation: Take a strong position; earn the strength; commit; rally.
- Reader stance: Either a recruit (open to being moved) or a skeptic (testing for sloganeering).
- Risks: Slogan without substance; substance without commitment; unearned grandeur.
- Strong move: Specific concrete claims that justify the manifesto's force; honest acknowledgment of what the position costs.
