Content Factory

Skill packs › Skill

Content Factory

Turn ONE recording into a full library — definitive article, clips, reels, posts, and ad-ready creative — through Dennis Yu’s factory line Produce → Process → Post → Promote, with every asset pointing at one canonical hub. Use when you have a recording to repurpose or a content system to design.

Skill file content-factory.md · last updated Aug 22, 2026

How to run it. Download any pack from the skill-pack directory, unzip it into your Claude project folder, and this skill is one of the files inside. You do not paste it anywhere — the agent reads it when the job calls for it.

Use this when you’ve captured one thing — a Zoom interview, a 60-second hallway clip, a conference talk — and want a content library that compounds instead of a post that disappears. Step 8 of the Local Service Spotlight method, and the engine the other nine skills feed.

Inputs

  • One raw recording (video or audio) + transcript. Phone quality is fine — Marko Sipila built HVAC Quote on phone-shot conference interviews.
  • The subject’s positioning and proof (from personal-brand-strategist and positive-mentions-harvester).
  • Your lighthouse list — the marquee people and brands you’re tied to.
  • Your canonical hub URLs, so nothing you publish competes with them.

The factory line — Produce, Process, Post, Promote

Run the line in order. Each station feeds the next; skip one and the next breaks.

  1. Produce — capture real source material: a recording, customer story, job, demonstration, review, or interview.
  2. Process — transcribe, verify, mine, edit, and atomize the source without inventing claims, quotes, or proof.
  3. Post — assemble the one canonical hub and its channel-ready derivatives. Stage by default; publish only under the caller’s exact action authority.
  4. Promote — rank proven organic performers and prepare Dollar-a-Day recommendations (dollar-a-day-strategist). Live spend, kill, scale, and audience changes require a scoped approval receipt.

Outside the factory (do not call these factory stages): Plumbing is onboarding / access / tracking before the line runs. Perform is MAA (Metrics → Analysis → Action) after promotion. Older material that lists Plumbing, Publish, Promote, Perform is a wrapper for the broader Social Amplification Engine — not a second name for this factory line.

Canon: Content Factory · The 4 Stages of The Content Factory · skill source dennisyu/local-service-spotlight-skills …/content-factory/SKILL.md.

Where the other skills plug in

  • Plumbing = client-access-checklist plus measurement contracts. Search Console, CMS, GA4, and Tag Manager must be verified green before shipping or amplification.
  • GCT = the accepted business brief from gct-screen, with targeting last.
  • Factory = the 6-stage per-recording line below (maps onto Produce → Process → Post → Promote).
  • Promote = dollar-a-day-strategist, recommendations by default.
  • Perform = the weekly weekly-brand-maa read (after the factory).

The assembly line — 6 stages per recording

  1. Capture — record once. One Zoom interview or one 60-second conference clip is a month of raw material.
  2. Transcribe & mine — pull the strongest quotes, claims, numbers, and stories out of the transcript.
  3. Definitive article — write the ONE canonical long-form hub (definitive-article-writer). The hub is the product; everything else is distribution.
  4. Atomize — cut clips, reels/shorts, and tweets from the same recording, Descript-style. One strong quote = one asset; an hour of recording yields 10–20 clips.
  5. Distribute — prepare channel-specific packages for YouTube/Shorts, LinkedIn, Instagram, TikTok, X, Facebook, and email. Publish/send only under exact action authority. Every approved asset links back to the hub; never create a competing page.
  6. Boost recommendation — prepare the $1/day test for proven winners (dollar-a-day-strategist). Do not spend without the exact approval scope.

The org chart — worker, manager, architect

  • Worker function — cuts clips, prepares channel packages, and reports against the checklist.
  • Manager — runs the workflow, enforces the SOP, reviews QA before anything publishes.
  • Architect — designs the system: sets the Goals in GCT, picks topics off the topic wheel, approves the definitive article, owns the standards.

You are the architect. Build the factory; don’t run the line. Stages 1 and 3 need you — your face, your facts. Stages 2, 4, 5, and 6 belong to workers. If you’re personally editing clips, the system failed — write the SOP that makes a worker’s output indistinguishable from yours.

First run — this week

  1. Book one 30-minute Zoom where someone interviews YOU about your differentiation from step 1.
  2. Run stages 2–5 on it; assign each stage to an owner function with a deadline.
  3. Stage the two strongest clips for a $1/day test. After a human approves the exact account, amount, and dates, read the MAA evidence at day 7. The agent recommends the decision; it does not autonomously spend, kill, or scale.

Output

  • One definitive article + a pack of clips, shorts, tweets, and ad-ready creative — all cross-linked to the hub.
  • A written SOP per stage, so a worker runs the next recording without you.
  • Cadence: capture monthly, atomize weekly, stage proven-winner tests, and review MAA weekly — adjust the system, not just the posts.

For DealCon — agency owners & acquirers

If you run an agency: a productized content factory — pre-defined offers, the Apple/Tesla factory model — lets you sell a system instead of hours, charge premium for repeatable results, and hand a buyer an agency that runs without you. If you buy & sell companies: install the factory in every acquisition post-close, so marketing stops being the founder’s personal hustle and becomes transferable SOPs — that shows up directly in EBITDA and the exit multiple. Your edge: name the one topic only you could talk about for 60 unscripted seconds. That’s your first capture — schedule it this week.

Run on a persistent agent (Fable 5)

  • Loop to done: one recording in, all six stages staged — the run is not done at the definitive article; loop until the clips are cut, channel packages are ready, every asset points back to the hub, and proven candidates are queued for dollar-a-day-strategist. External execution remains a separate approved state.
  • Self-verify: QA each stage against its SOP before the next stage starts — a deviation that ships compounds downstream.
  • Compound with memory: hold the hub URLs, lighthouse list, and per-stage SOPs in memory so every new recording drops into a running factory instead of a rebuilt one.
  • Log the run: per-stage notes become the SOP updates that make the next run cleaner — you’re the architect, and the log is how the factory improves.

See boil-the-ocean.md for the full operating principles.

Notes — Dennis’s method

  • The proof: Marko Sipila shot phone interviews at HVAC conferences, repurposed them onto the HVACQuote YouTube, boosted the winners at a dollar a day, and scaled HVAC Quote past 300 customers. Interview once → repurpose → boost.
  • Same three moves behind Harry Gold (harryjgold.com built from a single Zoom interview) and the Superior Fence & Rail footage with Zach Peyton (largest US fence franchise — 110+ locations, $310M+).
  • Feature your lighthouse. A clip with a marquee name converts harder than a clip alone — their authority reflects onto you.
  • Standardize and productize. Custom work doesn’t scale and doesn’t sell; a factory does. Boil the ocean on the system once, then let workers run it forever.
  • The factory is the Post-and-Promote muscle behind your entity home (step 4) and Knowledge Panel plumbing (step 5) — corroboration at scale.

Definitive article & pairings

Learned in the field

Appended automatically by the self-improvement loop (Skill-Learnings/): dated lessons from real runs. Newest at the bottom.

<!– learning:2026-07-20-daily-video-watchdog-loop –> July 20, 2026 (from: Absorption of Daniel Goodrich’s Repurposing Suite v0.2 (Basecamp message 9944086092) into the master pack)

The Capture stage now has a daily automated feeder: video-repurposing-agent (added to this pack July 20, 2026) watches the client’s YouTube channel every day with an incremental watermark (first run inventories the whole uploads playlist; later runs stop at the first already-seen video), triages each new upload SKIP / LIGHT / MODERATE / STRONG against the GCT, and routes the winners through definitive-article-writer → jennifer → publish. Two rules it enforces that this factory inherits: search the site BEFORE writing and enhance the existing article (embed the new video there) when one already owns the topic — never ship a competing page; and when a client runs both a personal-brand site and a company site, one video may become two different articles (first-person cut and third-person cut, cross-linked), never the same article twice. Proven economics from the Escape Fitness validation (May 2026, Daniel Goodrich): a 247-episode channel inventoried and the first A- article cost 251 of 10,000 free daily YouTube API units.


Other skills: ai-search-visibility · boil-the-ocean · client-access-checklist · client-relationship-cadence · content-agent · definitive-article-writer

The full run order is on the skill-pack directory. Every skill here is one task from the Task Library — the library is the catalogue of what can be done; a pack is the subset you install; an agent is who runs it.

Where this sits in the system

A skill is a document. A pack is a folder of documents. Neither one does any work. Work happens when a job runs those skills on a schedule, checks its own output, and keeps its files somewhere it can read them again tomorrow. That is the whole difference between owning skills and having an agent.

  1. Skill — one task, written down to a standard, so an agent can run it without you in the room. There are 239 of them.
  2. PackYOU ARE HERE — those skills bundled into a download you install in one paste.
  3. Agent — a named role with a job description — not a chat window you retype every morning.
  4. Job — a schedule, a QA cycle, and somewhere to keep working files. Miss any of the three and nothing runs twice.
  5. Proof — every finished run written up in public, and the lesson pushed back into the skill.

The map: The System · every asset: Asset Tracker · next door: Asset Tracker.

Scroll to Top