Memory Wiki

Good, I have all three inputs and the keyword-agent skill loaded. Let me analyze: **Script essence:** Claude Hopkins as a young bookkeeper at Bissell — the only person willing to check the numbers on a celebrated ad and speak truth to power. **NLP highlights:** Halden (boss, salience 0.199), Ethan (friend, 0.069), bookkeeper (0.07), Bissell Carpet Sweeper Company — the story is dramatized historical fiction, not a lecture. **Ideal avatar:** B1-C1 professionals (SEO managers, digital marketers, tech leads) — career-driven, no tourists/hobbyists/beginner content. Now let me generate candidates and validate the top ones with SERP.

Now I have enough validation data. Let me compile the complete structured keyword set.

Here is the complete structured keyword set. The full JSON has been saved to `/root/episode1-keywords.json`. --- ## What I Did **Step 1 — Read all three inputs:** - `/tmp/episode1-script.md` — full episode transcript (dramatized historical fiction: Claude Hopkins vs. a boss who hates data) - `/tmp/nlp-output.json` — 80+ entities extracted. Key signals: Halden (salience 0.199), Ethan (0.069), bookkeeper (0.07), Bissell Carpet Sweeper Company - `/tmp/ideal_avatar.md` — B1-C1 professionals: SEO managers, digital marketers, tech leads. Anti-tourist, anti-beginner, anti-academic **Step 2 — Generated 22 candidate keywords** across head terms (3), long-tail phrases (9), question-based (5), and topic clusters (4 groups × 4 keywords each) **Step 3 — Validated top 7 candidates via 7 web_search calls** against real Google SERP results. Key findings: | Search Query | Verdict | |---|---| | "Claude Hopkins" + "learn English" | **COMPLETE GAP** — zero ESL content exists | | "learn English through advertising stories" | **GAP** — no one targets this sub-niche | | "business English listening for marketers" | **MODERATE** — generic lesson plans exist, no narrative | | "history of advertising" English listening | **GAP** — lesson plans exist, no dramatized stories | | "scientific advertising" English vocabulary | **GAP** — ELL user asked for help with Hopkins' English | | "business English through stories" | **MODERATE** — StoryLearning/BBC but no advertising focus | | "speak confidently English meetings" | **WEAK MATCH** — dominated by courses, use as internal nav only | **Step 4 — Wrote structured JSON** to `/root/episode1-keywords.json` --- ## Key Findings **The biggest opportunity:** Claude Hopkins + English learning is a **complete blue ocean** — zero existing content, strong intersection of domain expertise (advertising history) and language acquisition. The ELL StackExchange user who asked for help understanding the English in Scientific Advertising proves real demand exists. **Primary keyword recommended:** `"learn English through advertising stories for B1 professionals"` — 8-word long-tail, zero competition, targets the ideal avatar specifically, naturally differentiates from all other episodes (which would use different domain terms like "sales stories," "copywriting stories," etc.). **Episode differentiator for the writer:** Episode 1 owns the **origin story** angle — a young man discovering scientific advertising through painful experience. The keywords `"data-driven decisions"` and `"speaking truth to power"` belong to this episode and should not be reused on later episodes. **JSON written to:** `/root/episode1-keywords.json` (12.7 KB, valid JSON — lint passed)