A Friend — Content Distribution System v1
Shared operating plan between Daniel (founder) and Daniel (content partner). Updated as formats prove themselves.
0. Two Pipelines — Do Not Conflate
| Pipeline | Owner | Purpose | Lives In |
|---|---|---|---|
| Compose Skill | Founder only | Raw thinking → narrowed writing for personal channels (Substack, X, LinkedIn) | Local codebase, Claude Code |
| Content System (this doc) | Founder + Daniel | A Friend production across Insta/TikTok, 7 posts/week target | Shared, lives here |
The pipelines cross-pollinate at one specific point (see §6), but their workflows, tools, and ownership are separate. This separation is the thing that has been collapsing under fragmentation. Hold it.
1. Operating Doctrine
- Cadence over polish. Algo rewards daily activity. 7 posts/week, low friction per post.
- B-roll bank is the moat. Every Daniel session feeds it. Most formats consume from it; only a few generate fresh footage.
- Daniel runs three formats unsupervised. Founder-led formats are the ones that need scripting overhead. Protect Daniel’s autonomous output — it’s where leverage lives.
- Scripts only where they buy something. Challenge format and Reaction format need scripts. The other four don’t.
- One backlog, not five. Ideas land in one place between sessions, get triaged at the start of each Daniel-session.
2. Format Inventory (6 Formats, 7 Posts/Week)
Tier A — Daniel-Autonomous (no scripting from founder)
A1. B-Roll Organic + UGC Scenarios
- What: lifestyle/atmospheric capture + real-world A Friend usage moments
- Production: Daniel shoots whenever he’s around the founder
- Cadence: 1 post/week, fresh footage feeds the bank for other formats
- Friction point: UGC scenarios sometimes blocked by Daniel not knowing A Friend’s current capabilities → see §7 (Capability Cheatsheet)
A2. “Best Friend’s Content Creator” POV
- What: Daniel’s own POV format. Raw founder-journey docu-style. His voice, his framing.
- Production: fully unsupervised, Daniel decides angle and timing
- Cadence: 1 post/week (candidate for 2x/week if Daniel has bandwidth)
- Why this matters: most authentic founder-adjacent content angle we have. No script will ever match its tone.
A3. Carousels
- What: slide-throughs from existing video/photo library, occasionally fresh stills
- Production: Daniel assembles from B-roll bank
- Cadence: 1–2 posts/week (the easy throughput format)
Tier A subtotal: 3–4 posts/week
Tier B — Founder-Led (needs scripting)
B1. 1M Challenge Format
- What: “1M revenue in 6 months with a product that doesn’t fully exist yet” — high-energy founder journey, progress updates, what shipped, what failed
- Production: founder scripts → captures with Daniel during his session → Daniel cuts
- Cadence: 1–2 posts/week (the most leverageable founder-led format — can grow without diluting)
- Weaves into Daniel’s B-roll sessions naturally. Default to this when uncertain what to script.
B2. Reaction / Educational
- What: AI/web advancements, founder commentary, thought-leadership clips
- Production: content extraction skill pulls sources → founder scripts/records → Daniel cuts
- Cadence: 1 post/week
- Slower production. Quality over volume. Skin-in-the-game angle.
Tier B subtotal: 2–3 posts/week
Tier C — Cross-Pollination (Personal ↔ A Friend)
C1. Collab-Posting (from Substack)
- What: Long-form essay on personal Substack → atomized to X/LinkedIn → where Onari/A Friend topics overlap, repurposed for A Friend channels too
- Production: founder Compose Skill pipeline (separate). Output: Substack post + tweets + LinkedIn posts. Sometimes a YouTube long-form once that pillar is live.
- Cadence: 1 post/week on A Friend channels (when topic overlap exists)
- This is the bridge between the personal voice and the product channels. Don’t force overlap — let it happen organically.
Tier C subtotal: ~1 post/week on A Friend channels
Weekly Math
| Format | Posts/wk | Owner |
|---|---|---|
| B-Roll/UGC (A1) | 1 | Daniel |
| BF Content Creator (A2) | 1–2 | Daniel |
| Carousels (A3) | 1–2 | Daniel |
| 1M Challenge (B1) | 1–2 | Founder + Daniel |
| Reaction/Educational (B2) | 1 | Founder + Daniel |
| Collab from Substack (C1) | 0–1 | Founder |
| Target | 7 |
If Daniel expands beyond 10 hrs/week: A2 or A3 doubles first (highest ROI).
3. Daniel’s Session Decision Tree
When Daniel arrives, run this in order:
- B-roll bank low? → first 30–60 min: organic capture + atmospheric shots. No script needed.
- Challenge update scripted for today? → capture it. Energetic, short, the founder is in-frame.
- UGC scenario queued? → execute (founder-scripted) OR Daniel improvises if A Friend capability is clear.
- Carousel idea in backlog? → Daniel assembles from existing library at the end of session or async.
- Reaction footage needed? → only if founder has the script ready. Otherwise defer.
- BF-Content-Creator → Daniel decides on his own timing; not a session-driven format.
Hard rule: Daniel never leaves a session without (a) fresh B-roll added to the bank, or (b) at least one cut deliverable in flight.
4. Time Allocation (Daniel’s 10 hrs/week)
Rough split. Adjust after 2 weeks of data.
| Activity | Hours | Notes |
|---|---|---|
| Shooting (B-roll, UGC, Challenge captures) | 3–4 | with founder present |
| Cutting Tier A formats (A1, A2, A3) | 4 | mostly async |
| Cutting Tier B formats (B1, B2) | 2 | with founder QA |
| Triage + planning (start of session) | 0.5 | the backlog review |
Founder time on the system: 3–5 hrs/week. Mostly scripting B1 + B2, plus running the Compose pipeline for C1.
5. What Needs Scripting vs. Not
| Format | Script? | Why |
|---|---|---|
| A1 B-Roll/UGC | No (UGC: light bullet sometimes) | Atmospheric, organic |
| A2 BF Content Creator | No | Daniel’s voice is the point |
| A3 Carousels | Light caption only | Visual carries it |
| B1 Challenge | Yes — short script per capture | Energy + framing matters |
| B2 Reaction | Yes — fuller script | Thought-leadership accuracy matters |
| C1 Collab from Substack | Already written | Compose skill output is the script |
6. Cross-Pollination Rules
The personal channel (Substack → X, LinkedIn) and A Friend channels share a substrate (Onari powers A Friend, so the philosophy overlaps). But they are not the same audience.
Overlap is allowed when:
- The Substack essay treats a topic A Friend’s audience would also want (e.g. “voice as the way humans actually think”)
- The atomization produces content that doesn’t require Onari technical context to understand
Overlap is NOT forced when:
- The Substack post is deeply about Onari architecture, GSD-2, dev-internal stuff → stays personal only
- The Substack post is about founder vulnerability or personal stakes → stays personal only (the BF-Content-Creator format already covers this angle for A Friend channels)
Mechanism: after each Compose Skill run, tag the output [personal-only] or [cross-eligible]. Cross-eligible pieces become A Friend posts within the same week.
7. Open Build Items (Required for System to Run)
These are blocking or near-blocking for the system to function at 7/week consistently:
- A Friend Capability Cheatsheet — single-page doc, updated weekly when features ship. Daniel reads it before any UGC capture. Eliminates the “lost on capabilities” friction. Owner: founder. ETA: this week.
- Backlog System — one place where ideas, scripts in-progress, B-roll prompts, and reaction-source-links land between sessions. Could be a Notion page, a Reminders list, or a markdown file synced via GSD-2. Pick the lowest-friction option. Don’t over-tool.
- Content Pool / Extraction Skill — revive — already exists. Restore as the source feeder for B2 Reaction format. Owner: founder. ETA: 2 weeks.
- Compose Skill — refresh — exists, dusty. Reactivate for C1. First run: pick one Substack essay from the transcript backlog and atomize it. Owner: founder. ETA: 1 week.
- Posting calendar — light Notion/Sheets view: what posted when, what’s queued. Visibility for both Daniels. Not analytics — just queue management. Owner: Daniel-content. ETA: this week.
8. What Success Looks Like at 4 Weeks
- 7 posts/week sustained for at least 3 of the 4 weeks
- B-roll bank doubled in size
- Capability Cheatsheet exists and Daniel has used it to make at least one UGC piece autonomously
- At least 2 Substack essays published and atomized; at least 1 cross-eligible piece ran on A Friend channels
- The “what do I do today?” friction is gone — every day has a default format queued
If we’re at this state at week 4, the system works. If not, we cut formats, not increase friction.
9. What This Doc Is NOT
- Not an editorial guide (tone, brand voice, hooks — separate doc)
- Not an analytics review (what’s working — that’s a weekly retro, not this)
- Not the Compose Skill spec (that’s the personal pipeline, separate)
- Not a substitute for instinct — if a format starts feeling forced, kill it
Version: v1 — initial capture Review cadence: weekly retro; major revisions every 4 weeks Last updated: 2026-05-15