12.3 Scaling to a Portfolio of Personas

A second persona is not diversification. It is duplication with a shared cost base — and if the first persona is not yet profitable and self-running, the second one mostly accelerates the failure of both.

What can be shared, and at what quality cost

The rule that follows: the infrastructure is shared, the identity is not. Every attempt to share identity — one face on two accounts, two names for one character — produces either a duplicate account, which platforms treat as spam, or a confusing one, which audiences treat as a mistake.

Why the same aesthetic on two accounts hurts both

The counterintuitive one, and it is not a quality argument.

Reach is partly aesthetic. A grade, a palette, and a framing habit are what make an account recognizable in a feed, and recognition is a distribution mechanism (8.2). Two accounts with the same visual signature read to the platform as one account with two profiles and to the audience as a repost — which is the failure mode you least want, because it makes both accounts look derivative of each other.

Recommendation systems separate accounts that look different and merge accounts that look similar. The practical consequence: identical aesthetics compete for the same audience segment, and the second account's growth comes partly at the first's expense.

What to differentiate instead of everything. Three things, and no more: grade, subject presence (framing and wardrobe register), and voice. Those three establish that these are different characters, and the rest of the infrastructure can stay shared. This is cheaper than building a parallel pipeline and it is what makes the two accounts additive rather than cannibalizing.

The marginal economics

The conclusion, with a number attached: the second persona costs roughly sixty to seventy percent of the first, not fifty percent, and the savings come from infrastructure rather than from identity or from attention.

And the revenue does not arrive proportionally faster. The second account starts with no audience, no account history, and no algorithm learning, which is the same cold start as chapter 8.7 — with better tooling. A portfolio of three personas in year one is usually three accounts in the low hundreds per month rather than one account at three times the revenue.

Portfolio management

Once there is more than one persona, the decision that matters is which to keep, and it needs a rule rather than a feeling.

Revenue per hour is the ranking column. It is the only figure that compares a mature account against a new one on the same basis, and it is the one that makes the decision obvious: B and C together produce less than a third of A while consuming more hours than A.

The exit rule: a persona below a defined revenue-per-hour floor for two consecutive quarters goes to the wind-down procedure in 12.6. Define the floor before you need it, because once you have an account with an audience and a handful of regular subscribers, the decision stops being analytical.

And the rule that prevents the trap: never launch a new persona while any existing persona is below the floor. A portfolio grows by fixing and retiring, not only by adding. [UNVERIFIED] The revenue-per-hour figures in the table are illustrative composites, not measured accounts; the ranking method is the part that transfers.

The handoff that makes a second persona possible

The second persona is only cheap if the first one's process is written down. Four documents have to exist before the gate opens, and each of them is a chapter from this book in finished form.

The test, and it is a good one: could a competent stranger stand up persona two from these four documents plus a new face? If not, the gap is the thing to write before the second persona rather than during it.

Why this doubles as value for persona one. A process documented well enough to be handed over is a process that can be delegated (10.5), audited (12.5), and sold (12.6). The second persona's prerequisite is the first persona's exit readiness, which is why writing it down is never wasted work.

Designing persona two differently

Four decisions to make at design time rather than after launch, because all four are expensive to change later.

The differentiation set. Grade, subject presence, and voice — the three from earlier in this chapter. Decide them explicitly and write them into the new bible as constraints, not preferences.

The audience overlap. Choose a niche adjacent enough to reuse your category knowledge and separated enough that the two accounts do not compete for the same followers. The test: would a follower of persona one plausibly follow persona two? If yes for most of them, the overlap is a cannibalization risk rather than a cross-promotion opportunity.

The platform split. Two personas on the same platform, posting in the same category, at the same cadence is the configuration most likely to trip duplicate-account enforcement (8.5). Decide whether persona two leads on the secondary platform or on a different one entirely.

The cost accounting. Assign shared infrastructure cost to persona one and charge persona two only its marginal cost (12.2). This makes persona two look cheap, which is deliberate: the question you want the numbers to answer is whether the marginal venture pays for itself, not whether it covers a share of something that already exists.

Where this goes wrong

Launching the second before the first is self-running. Two neglected accounts.

Reading one good month as a green light. The gate requires two, deliberately.

Sharing a face or a name. Duplicate or confusing; platforms and audiences both notice.

Identical aesthetics on both accounts. They compete for the same audience.

Expecting a 50% cost. It is 60–70%, and the savings are not in the expensive parts.

A portfolio with no retirement rule. The weakest account consumes hours indefinitely.

Adding a persona during a decline on the first. Attention is the constrained resource.

Judging personas by revenue alone. Revenue per hour is the comparable figure.

The gate

1. Has the first persona earned above $1,000 a month for two consecutive months?
2. Is its recurring work delegated or systematically reduced?
3. Can its pipeline run for a week without your attention?
4. Have you differentiated grade, subject presence, and voice rather than everything?
5. Do you know your marginal cost as a percentage of the first persona's?
6. Do you have a revenue-per-hour floor and a retirement rule written down before you need them?

What you leave with

A three-condition gate for the second persona with the test that fails most often, a sharing table that separates infrastructure from identity with the rule that follows, the argument that identical aesthetics cannibalize reach and the three things to differentiate instead, a marginal-cost table showing 60–70% rather than 50% with the cold-start warning, and a portfolio table ranked by revenue per hour with an exit rule and the rule against adding while anything is below the floor.