Own posting
Posting only on accounts in Ralph's own name, to build the hook-outcome dataset. Never a farm.
Verdict: optional. Not on the critical path.
Live state: Mission Control, every clip, account, queued post and prospect, rebuilt by the loop twice a day.
The machine is real. The customer is not yet. The business is cut, caption, hook, schedule on the client's own accounts. It earns the right to exist with a 14-day demand test that costs Ralph about one hour.
Posting only on accounts in Ralph's own name, to build the hook-outcome dataset. Never a farm.
Verdict: optional. Not on the critical path.
Cut, caption, hook and schedule to the creator's own TikTok, Reels and Shorts, through a scheduler they OAuth into. Original content stays on the original account.
Verdict: the business. 3 to 5 clients is the honest ceiling for Ralph's calendar.
A brand funds a clipper pool. Ralph would recruit, vet, approve and report, and keep a fee on the budget.
Verdict: parked until a signed client asks. No brand has asked and none is named on disk.
This rerun includes 30% churn every 90 days, Whop's 10% fee, scheduler cost instead of editor cost, and no Layer C.
The modal outcome of cold outreach, and the old model had no row for it.
Two or three retainers if the first calls go slowly. Still real money, still on the client's own accounts.
Three to five retainers on about five hours a week, real money, not the $17.8k that was on this page this morning.
Only if the test converts and Ralph wants the bigger version. Layer C appears only here.
Hard kill date: day 14. If it fails, the engine becomes an internal content tool and this venture closes.
Goal: every verified name has a locally-cut pack, and nothing is hosted.
Goal: bad clips killed, one two-line email sent per verified name.
Goal: 2 yeses to the clips and 1 creator willing to pay, or the venture closes.
The modal outcome of cold outreach. The old model had no scenario for it.
GuardThe 14-day test exists to find this out for one hour of Ralph's time, not three months.
Cali Life Q4, RWP, Helm + Oar, the job search, Greenvale. The plan's own words: hours are the constraint.
GuardHard cap 5 hours a week. Cut order when time runs out: own posting first, dashboard polish, then reporting. Never cut pack QC or the reply-within-the-hour rule.
It reported visual confirmation on 35 rows it did not see.
GuardNothing enters outreach unless a script verified it against live YouTube. Structural, not a reminder.
Ralph's own standing rule forbids it, and 15 cold sends a day from gmail.com with a pages.dev link is a spam-folder strategy.
Guardsend.py refuses any gmail.com sender. A domain mailbox with SPF and DKIM is the next build.
The six survivors may be survivors because Shorts do not sell their product. The warmest buyers may be the ones posting 14 Shorts that go nowhere.
GuardA second pitch track, UNDERPERFORMING: posts Shorts, weak median views versus long-form. Your cut versus ours. Being added to the verifier.
Unchanged. Marketplace balances get held for months.
GuardClient invoices are the revenue. Own posting is a dataset, not income.
Five adversarial advisors, a blind peer review, a chairman ruling, and every claim checked on disk by Ralph before it was accepted.
Instagram removes accounts that mainly repost from recommendations, judged at account level over a rolling 30-day window. Speed changes and watermarks do not count as original. A faceless network is an aggregator by definition, whatever the edit. It also costs Ralph about 16 hours of phone warm-up he does not have, and it answers no demand question.
That is a $4,500 ask of a stranger with no case study. First three clients are $1,000/mo, month to month, 30 clips or free, tested on a call, never in an email.
No brand has asked and none is named anywhere on disk. Removed from every scenario but Stretch until a signed client asks.
Zero churn, $400 per client of editor cost the engine already replaced, Whop's 10% fee missing, and a client booked in month 2 with no outreach sent.
Cali Life has no long-form to clip. It is not a test bed.
It contradicted the plan's own rule to license before clipping a frame, and put derivative cuts of finance creators' videos on a public URL under Ralph's account. Packs are now cut locally, and nothing is hosted or sent until the creator says yes.
Base is now 3 to 5 retainers on the client's own accounts, $3k to $6k a month net, about 5 hours a week. Same word, honest number.
Highest CPM in the market and the voice Ralph already writes in. Gambling and crypto-casino pools are blocked in code.
The engine removed most of the reason to hire. Editing, reframing, captioning and scheduling are now local and free per clip. Contract editors are only needed for QC at volume, a month-4 question, not a month-1 one.
15 minutes: watch the pack clips on the dashboard and kill any with a caption over on-screen text or a dead hook. 5 minutes: read the two-line permission-first email and write the approval file. No link, no clips, no price in it.
Both steps happen before anything else Monday morning.