The Problem With Content Teams
Most companies produce content in drip mode: write one piece, publish it, write the next. This approach kills productivity because context switching is expensive. You pick a topic, research it, write the first draft, edit, publish—then start over.
The better approach is batch production. But nobody teaches this.
We produce 200+ pieces monthly (newsletters, blogs, LinkedIn posts, X threads, emails) with 2 people. Zero outsourcing. Zero agencies. Here's exactly how.
The 4-Stage Pipeline: Scout → Analyst → Creator → Director
Stage 1: Scout (5 hours/week)
Scout monitors 50+ sources for patterns: industry newsletters, Twitter, Reddit, Hacker News, competitor blogs, research papers. When 3+ sources mention the same topic, that's a signal.
Output: 20–30 story ideas daily. We keep the top 5 for deeper analysis.
Stage 2: Analyst (8 hours/week)
Take the top 5 ideas from Scout. Deep research: read full articles, extract key takeaways, find the nuance. Build editorial context. Why does this matter? Who should care? What angle works for our audience?
Generate 4–5 angles per story:
- Thought leadership take
- How-to / practical angle
- Contrarian hot take
- Beginner-friendly breakdown
- Connection to industry trends
Output: 20–25 content angles daily. This is where editorial voice lives.
Stage 3: Creator (12 hours/week)
Write all first drafts using Claude API. In parallel:
- Newsletter: ~800 words, research-backed, data-driven, Guerilla Authority tone
- Blog post: ~1200 words, SEO-optimized, technical depth, includes resources
- LinkedIn posts: 4 pieces (150–250 words each), educational + shareable
- X threads: 4 threads (4–5 tweets each), hot takes + stat drops
- Email sequences: 2–3 per week, value-driven, soft sales, no hype
Cost: $0.30 per piece (Claude API). Time: 12 hours generates 1 newsletter + 1 blog + 4 LinkedIn + 4 X threads + 2 emails (12 pieces total).
Stage 4: Director (5 hours/week)
QA pass on every piece. Grammar, tone, accuracy, brand voice consistency. Fact-check stats. Tag metadata. Score submission.
Scoring system: 10 dimensions (1–10 scale each). 80+ = publish. <-80 = rewrite.
Last month: 124 pieces submitted. Average score: 91/100. Zero shipped below 80.
The Repurposing Architecture: 1 Newsletter → 12 Content Pieces
Here's where batch production becomes powerful. One source of truth generates 12 content vectors.
Input: One newsletter topic
Output:
- Newsletter (deep dive, 8–10 min read)
- Blog post (technical, SEO, 12–15 min read)
- LinkedIn Post A (executive summary)
- LinkedIn Post B (architecture breakdown)
- LinkedIn Post C (contrarian angle)
- LinkedIn Post D (personal story)
- X Thread A (quick hot take)
- X Thread B (deep dive technical)
- X Thread C (stat drop)
- X Thread D (contrarian take)
- Email sequence part 1 (problem statement)
- Email sequence part 2 (solution + CTA)
All generated in parallel. All maintained consistently.
Quality Gates: The Scoring System
We don't ship below 80. This single decision eliminates 90% of content problems.
| Dimension | What We Check | Penalty |
|---|---|---|
| Accuracy | Facts correct? Stats cited? Sources verified? | -5 points |
| Tone | Is this Guerilla Authority? Confident, data-driven, irreverent? | -4 points |
| Usefulness | Does reader learn something? Is it actionable? | -5 points |
| Structure | Does it flow? Clear headers? Scannable? | -3 points |
| Brand voice | Does it sound like us? | -3 points |
| SEO (blogs) | Keywords, headers, internal links? | -5 points |
| CTA clarity | Is next action obvious? | -2 points |
| Grammar | Typos? Run-ons? Consistency? | -2 points |
Real March Numbers
| Content Type | Volume | Total Words |
|---|---|---|
| Newsletters | 4/month | 3,200 |
| Blog posts | 4/month | 4,800 |
| LinkedIn posts | 16/month | 3,200 |
| X threads | 16/month | 1,280 |
| Email sequences | 8/month | 1,200 |
| TOTAL | 48+ pieces | 13,680 words |
Cost: $300/month (Claude API). $0 infrastructure. Zero team overhead.
Why This Works (And Why It Might Not For You)
This system works because:
- We own the brand voice — consistency is automatic when one person writes most content
- We set editorial standards — no approval chains, no dilution
- We use Claude API — cost is negligible
- We batch ruthlessly — context switching is the enemy
- We automate repurposing — one idea becomes 12 content pieces
It won't work if you:
- Outsource writing (inconsistent voice, slow iteration)
- Require legal review (adds weeks per piece)
- Have strict approval chains (kills batch efficiency)
- Need 500+ pieces/month (diminishing returns)
What's Next
Stage 5: Distribution Automation. Scheduled Slack posting, calendar integration, cross-posting to Reddit/HN/newsletter communities. Expected: +30% reach per piece.
The content machine doesn't stop improving.
Tools Mentioned in This Post
Some links are affiliate links. We only recommend tools we actually use.
- n8n — Workflow automation platform for content pipeline orchestration
- Claude API — Production content generation, ~$0.30 per piece
- Beehiiv — Newsletter platform, hosting and analytics
- Mac Mini M4 Pro — Local development + testing infrastructure
- MEWR Tools — Content optimization tools built on this architecture