content-strategy

Content Strategy

Every piece must be searchable, shareable, or both, in that priority order, because search traffic compounds and social traffic evaporates. Topics come from customer evidence, not brainstorms.

Workflow

1. Load context. Read any product marketing context file (.agents/product-marketing.md or .claude/product-marketing.md) first. Then gather what is missing: what the company does, ICP, content goal (traffic, leads, authority), current content and resources, main competitors.
2. Mine evidence for topics. Work the sources in references/ideation.md: keyword exports, sales and support call transcripts, survey responses, forum threads (Reddit, Quora, communities), competitor content gaps, and customer-facing teams. Every topic idea should carry its evidence ("asked in 6 of the last 10 sales calls").
3. Classify each idea: searchable, shareable, or both.
4. Searchable captures existing demand: target one keyword, match intent exactly, structure headings around search patterns, cover the question fully.
5. Shareable creates demand: novel insight, original data, a well-argued contrarian take, a story with stakes.
6. Define 3-5 pillars. The core topics the brand will own. Test each pillar: aligns with the product, matters to the ICP, has search or social interest, is broad enough to spawn 10+ subtopics. Map clusters under each pillar. Default all content to /blog with internal linking, dedicated hub URL structures only for major multi-layer resources.
7. Map topics to buyer stage using the modifier table in references/ideation.md (awareness "what is", consideration "best/vs", decision "pricing/reviews", implementation "templates/how to use").
8. Score and rank. Customer impact 40%, content-market fit 30%, search potential 20%, resource requirements 10%. Output a ranked table.
9. Deliver. Pillars with rationale, prioritized topic list (each with type, target keyword, buyer stage, and evidence), and a cluster map showing how pieces interlink.

Verification

Do the evidence audit: for each of the top 10 recommended topics, expect a named evidence source (transcript quote, keyword data row, forum thread, support ticket pattern). Count how many rest only on "seems like a good topic". More than zero: go back to step 2 for those, or move them to a clearly labeled speculative list.

Good vs Bad

Bad: "You should start a blog about productivity, AI trends, and company culture." No evidence, no connection to the product, no keyword or persona attached.

Good: "Pillar: client onboarding. Evidence: 'how do I get clients to fill out forms' appears in 8 of 20 sales calls and maps to a 1.9k/mo keyword cluster your competitors rank weakly for. First three spokes: intake-form templates (implementation intent, product-adjacent), 'client onboarding checklist' (awareness), 'best client portals' comparison (consideration)."

Footguns

• Strategy by brainstorm. Topics invented in a meeting room rarely match what customers search or share. Fix: every topic carries an evidence tag.
• All awareness, no money pages. A blog of "what is X" posts builds traffic that never converts. Fix: cover all four buyer stages, weight consideration and decision for lead goals.
• Premature hub/spoke URL architecture. Custom /guides/ structures for a 10-post blog add migration pain for zero gain. Fix: /blog/post-title until a topic proves deep enough.
• Copying competitor calendars. Their strategy fits their positioning and authority, not yours. Fix: use competitor analysis to find gaps, not to mirror.

Completion checklist

• [ ] Context loaded, goal and ICP confirmed
• [ ] Topics mined from at least two evidence sources
• [ ] 3-5 pillars defined and tested against the four criteria
• [ ] Every top-10 topic has evidence, type, keyword, and buyer stage
• [ ] Scored ranking table and cluster map delivered

Any box unchecked: not done. Fix or say so.