12.1 The Operating Model for a Single Persona
Part VII described a pipeline. This chapter describes the job that operates it: how many hours a week it actually takes, which tasks never go away, and the point at which one person stops being enough.
The number matters because it is the input to every other decision in this part. Chapter 12.3's question about a second persona, 12.5's audit, and 12.6's exit all depend on knowing what the first one costs in time.
The week
The slot layout belongs to 7.6. What this chapter adds is what the slots cost: ninety, ninety and sixty minutes of production, a forty-five-minute review on Sunday, and a chat window that is not optional (9.4).
Production and review: about five hours a week. That is the number most operators underestimate on the way in and the one that determines whether the schedule survives.
Chat: three and a half to ten and a half hours a week, and it scales with the audience rather than with the posting volume. At a hundred paying subscribers a daily half hour is a minimum. At a thousand it is a job.
Strategy and admin: two to four hours a week — the quarterly audit amortized, price checks, policy reading, bookkeeping, and the tax and compliance overhead that 11.7 and 11.8 create for an EU operator.
The realistic total for a working single-persona account: ten to twenty hours a week. Not two, and not forty. The word "working" is doing load-bearing work in that sentence: an account producing three posts a week with no monetization runs closer to six hours.
This is a part-time business at one persona and a full-time job at two or three. Plan the schedule before the audience, because the audience is what creates the schedule.
Which tasks never go away
Two lists, and the split is the most useful thing in this chapter.
One-off tasks. Identity design (Part IV). Reference libraries (5.3). The dataset (4.4). Adapter training (4.5). The A saved generation workflow expressed as connected nodes, runnable from an API or script. (7.2). The account setup and the silence phase (8.2). The SOP (7.6). Each of these is finished, and each of them is also maintenance on a slow cycle rather than a permanent job — the adapter needs retraining when the face drifts (4.7), the library needs a quarterly audit, the SOP needs versioning.
Recurring tasks, forever. Generation and QA. The maintained set of forty reusable prompt variants, refilled in the weekly review slot. refill. Publishing. Comment replies and chat. Analytics reading. Disclosure compliance on every post. Cost tracking. Policy re-reading.
The operational consequence. The recurring list is what the schedule has to be designed around, and it is the list that does not shrink when the pipeline gets better — chapter 7.5's automation reduces the hours inside each task, not the number of tasks. An operator who plans for the one-off list and is surprised by the recurring one has mis-scoped the business rather than the tooling.
Roles at one, two, and three people
One person, everything. Viable while chat is under roughly ninety minutes a day. The failure mode is not capacity; it is that the same person doing production and conversation has no distance from either, and quality drifts in both.
Two people: producer and operator. The cleanest split, and the first hire. One person runs the pipeline — generation, QA, editing, scheduling. The other runs the audience — chat, comments, community, and the analytics that feed the prompt list. Chapter 10.5 covers the handoff mechanics, including the conduct rules and the compliance duty that stays with you regardless of who is typing.
Three people: producer, operator, and growth. The third role is distribution: channel work (8.4), the funnel (10.1), brand outreach (9.5), and paid testing. This is the first role that is not overhead, because it is the only one that adds reach rather than maintaining it.
| Headcount | Cost shape | What breaks first |
|---|---|---|
| 1 | Time only | Chat volume; no distance from the work |
| 2 | One contractor | Coordination; the voice drifting between two writers |
| 3 | Two contractors | Margin; three people is a business, not a project |
What never gets delegated: the publication decision (7.3), the disclosure posture, the price, and anything that commits the persona to a brand. Those four stay with the owner at every headcount.
Where the single-person model stops working
Three specific thresholds, in the order they arrive.
Chat over ninety minutes a day. The first limit. It is not that the hours are unavailable; it is that the time comes out of production, and production falling behind reduces the audience that feeds the chat. The loop is self-reinforcing in the wrong direction.
More than one post a day at video-first cadence. At 7.1's time table a video post costs twenty-five to thirty-five minutes of attention. Daily video is a full-time production job before chat is counted.
A second persona. Chapter 12.3 covers the decision; the operational fact here is that the second persona does not take half the time of the first. It takes roughly sixty to seventy percent, because identity work is new (Part IV) while infrastructure is shared.
The honest version of this chapter's conclusion: this business is a trade of time for revenue, and the exchange rate is worse than the marketing suggests and better than doing nothing. Chapter 9.3 covers what it is not. What it is, at one persona, is a part-time job with an unusually long ramp and an unusually low fixed cost.
The first ninety days
The hours do not start where they settle, and the shape of the ramp is worth knowing before it is experienced.
| Period | Weekly hours | What dominates |
|---|---|---|
| Month 1 | 15–25 | Identity design, dataset, adapter training, account setup |
| Month 2–3 | 12–20 | Production, silence phase, comment work, first chat |
| Month 4–12 | 10–20 | Steady state, with chat growing and production becoming routine |
Why month one is the heaviest. Part IV is front-loaded and it is the least compressible work in the book: the dataset must exist before the adapter, and the adapter must be validated before production means anything. [UNVERIFIED] Operators report that the identity phase takes two to three weeks of evenings, which matches the components rather than beating them.
Why the steady state is not lower than month two. Production gets faster and chat gets slower, and the two roughly cancel for the first year. This is the sentence that most surprises people who expected automation to buy them time; 12.2's per-persona sheet is where the trend becomes visible rather than assumed.
Where this goes wrong
Planning two hours a week. The schedule collapses in month two.
Treating chat as optional. It is where two-thirds of the revenue is (9.2).
Expecting the recurring list to shrink. Automation makes tasks faster, not fewer.
Hiring a chatter before the conduct rules exist. 10.5's handoff, skipped because it is urgent.
Hiring growth before production is reliable. Growth into a pipeline that cannot feed it.
Delegating the publication decision. The one irreversible step (7.3).
Adding a persona to solve a time problem. It multiplies the recurring list.
Measuring hours at month one. The steady state is higher, because the audience is larger.
The gate
- Do you know your weekly hours split across production, chat, and strategy?
- Can you name which of your tasks are one-off and which recur forever?
- Is chat under ninety minutes a day, or is it already competing with production?
- At your current headcount, which of the three thresholds is closest?
- Have you kept the publication decision, disclosure posture, price, and brand commitments with the owner?
What you leave with
A five-slot week with hours attached and a ten-to-twenty-hour realistic total for a working account, the split between tasks that finish and tasks that never do with the operational consequence that automation shortens tasks rather than removing them, role definitions at one, two, and three people with the four things that never get delegated, and three specific thresholds that mark the end of the single-person model.