LinkedIn Post Engine

Create LinkedIn posts that feel human, useful, and credible: proven hook frameworks, clear narrative structures, proof-first writing, and practical CTA design.

Voice calibration: when the context is Oded's voice or senior engineering leadership topics, read references/oded-voice.md and apply it — modes, formatting, Hebrew register, and judging criteria all live there.

Process Workflow

Phase 1: Audience + Positioning (CRITICAL)

Before writing, define:

1. Audience — Who is this for? (founders, recruiters, engineers, operators, marketers, sales leaders, etc.)
2. Goal — Reach, authority, leads, hiring, trust, replies, profile visits?
3. Core insight — What is the one thing worth remembering?
4. Proof — What makes this believable (numbers, before/after, constraints, mistakes, outcomes)?

If proof is missing, use placeholders like [X%], [Y hours] and request exact values.

When the user gives only a raw idea, ask/fill: Audience, Goal, Topic, Proof points, Tone, CTA preference.

Phase 2: Structure + Hook Selection

Choose a format first (Story, Framework, Contrarian, Case study, Teardown, Build-in-public), then select 2-3 hooks and finalize one.

Phase 3: Draft + Polish

• Keep paragraphs short (1-2 lines)
• Front-load specifics
• Remove generic filler; key claims and advice include specific reasoning or mechanisms, not surface statements
• End with one clear CTA
• Add 3-6 hashtags max

LinkedIn Feed Dynamics (Practical)

What usually performs best:

1. Strong first two lines (stops the scroll)
2. Specificity (numbers > adjectives)
3. Authentic pragmatism — willingness to discuss real constraints, tradeoffs, and what actually broke at scale (avoid classic "vulnerability" framing and over-used formulas)
4. Clear structure (easy to skim on mobile)
5. Conversation CTA (quality comments over empty likes)

Avoid: buzzword soup, generic motivation posts, overlong hashtag blocks, fake certainty without evidence, over-used LinkedIn tropes and boilerplate phrasing.

Hook Formulas

Important: Avoid generic/boilerplate hooks and over-used formulas (dramatic "everyone is doing X while they should be doing Y", "you won't believe...", overly polished contrarian openers). Prioritize hooks that sound like a real, pragmatic, experienced practitioner — leverage, constraints, clarity, systems-level observations.

The Contrarian / Leverage Hook (preferred style)

• "The highest-leverage move most orgs are still avoiding..."
• "Most teams optimize for local [metric]. The real constraint is almost always [system-level thing]."
• "We improved this by removing options and forcing clarity, not by adding process or tools."

The Specific Result Hook

• "In [timeframe], we improved [metric] by [number]. Here's how."
• "We cut [cost/time] by [X%]. Not with a new tool—by changing this one workflow."

The Mistake / Reality Hook (use sparingly, make personal)

• "I made this [role]-mistake for months. It cost us [outcome]."
• "We shipped the wrong thing fast. Here's what actually fixed it at scale."

The Framework Hook

• "The [3-step/4-step] framework I use for [outcome]."
• "If I had to restart as [role], I'd follow these 5 rules."

The Question Hook

• "Would you let your team do [X] without [Y]?"
• "What's your biggest bottleneck in [domain] right now?"

High-Performing Post Formats

1) "This runs now" (Operational Story)

Best for real systems, workflows, and automations.

Template: bold claim → "Not as a demo. As an actual [responsibility/workflow]." → "Here's what it does:" with 4-6 concrete bullets → measurable result → perspective line + CTA question.

2) Case Study

Template: starting problem → constraints → intervention → before/after metrics → key lesson → optional "comment TEMPLATE" CTA.

3) Contrarian Opinion (Hybrid with Framework elements)

Best for posts that challenge dominant narratives while providing structured insight.

Template:

• Open with dominant narrative + mild critique ("The dominant AI narratives right now are depressingly predictable...")
• Pivot sentence to what deserves more attention
• Personal grounding ("I've been using the staff archetypes framework from Will Larson's book for years...")
• Breakdown of implications (by archetype/role or key dimensions)
• Synthesis: overall observation on incentives, tradeoffs, or long-term effects + forward-looking implication
• Understated close + peer question CTA ("What are you seeing in your org?")

4) Framework Post

Template: name framework → 3-5 steps → one mistake to avoid → one practical "do this today" action.

Bullets should reflect genuine insight density. If a point is thin or generic, deepen the reasoning or remove it — do not pad for length. When the post also has a contrarian angle, blend in the Contrarian Opinion flow above.

5) Build-in-Public Update

Template: what shipped this week → what worked → what broke → what changed next → ask for informed feedback.

More format skeletons: references/post-formats.md.

Persuasion Principles

• Specificity: "Saved 3.2 hours/week" beats "saved time."
• Mechanism: Explain how, not just outcomes.
• Credibility: Mention tradeoffs, not only wins.
• Relevance: Tie insight to audience reality.
• Clarity: One post = one core idea.
• Earned Insight: Advice and observations must feel earned — explain why something works (or fails), surface real tradeoffs or second-order effects, use specific language over platitudes. Applies to substance regardless of post length or format.

Writing Rules

• Hook in first 1-2 lines
• 1-2 sentence paragraphs
• Plain language over hype
• Emojis sparingly, as section markers at most
• Keep claims realistic
• Never invent outcomes, clients, or credentials
• Hashtags: 3-6 max, niche + function + audience mix, at the very end

Output Contract

By default, provide:

1. 3 hook options
2. 1 full post
3. 1 spicier variant
4. 3 first-comment ideas (to deepen discussion)

Optional on request: Hebrew / English / mixed variants, X/Twitter adaptation, short + long versions, carousel text outline.

Common Pitfalls to Avoid

• Generic "AI changed everything" without concrete examples
• No proof signals
• Too many ideas in one post
• CTA mismatch (asking for leads on a pure thought-leadership post)
• Over-formatting with noisy symbols
• AI/org transformation posts that ignore role-specific downsides or long-term incentive effects on talent

Execution Checklist

Before finalizing:

• Hook is strong and specific
• Core claim is clear
• Includes at least one proof signal
• Easy to skim on mobile
• CTA invites real conversation
• Hashtags are relevant and limited
• Tone feels authentic to author (voice-calibrated posts: run the judging criteria in references/oded-voice.md)