← Governance

NaughtyBits — Business Plan

The commercial structure · how the money is built to serve the constitution, not despite it · v0.1 draft, 2026-06-14

Status — read first. This is not law. The law is naughtybits-core.md; this is the commercial structure that must satisfy it, the way Forge and Community must. Where this plan implies a change to the constitution — specifically §9's "form — the lock," which currently names a stichting that owns the operating entity — that change is proposed, not made. It touches the ordinary, amendable layer (§8: the legal vehicle is explicitly meant to be worked out, and ../TASKS.md carries it as open), but it must still satisfy the entrenched guarantees of §9 — non-suppressible, non-defundable, a finding the entity cannot kill. This plan is written to satisfy them, and §6 below shows its work. Nothing here is ratified.

It is also honest about its largest gap up front: the specific commercial products that generate revenue are not yet decided. What is decided, and what this document fixes, is the architecture of the money — how whatever revenue arrives is governed, shared, and prevented from corrupting the oversight that constrains it. A business plan that invented a product line to look complete would be the apologetics-with-footnotes §4 forbids. So §3 defines what revenue is even permissible here, and §8 marks the product choice as the open decision it is.

1. The thesis: ethics-before-profit, made structural

Every venture in this domain claims to care. The claim is worthless as a promise and is only worth anything as a structure — something true about where the money and the control sit that would survive the founder losing his nerve or his virtue. This plan makes "ethics before profit" structural in three moves:

1. The profit stream is divided evenly across the parties — the affected-persons organisations and the founder, as equal partners. The population whose domain this is holds a real, non-revocable stake in the value created in it; the founder is one partner among them, aiming for a living, not a fortune — not a majority owner who shares some down. An equal entitlement, not a donation.
2. Control and the kill-switch sit in a vehicle the entity cannot capture — separated from the people who profit, so the platform's success can never buy its own oversight's silence.
3. The oversight body's funding is equity, not a budget — which fixes the single failure §9's own research identified in the Meta Oversight Board: year-to-year funding left a slow lever to starve it. A share of the profit stream, held as a non-clawback-able certificate, is far harder to withdraw than a line item.

The test of all three: doing right by the population is entangled with the entity's existence, and the people who can stop the business are not the people who profit from it.

2. The entity and its vehicles

A commercial venture from the start — not a charity in a nonprofit costume. The structure separates three things conventional ownership fuses: operation, control, and economics.

This is the crux: the STAK separates control from economics. It holds the legal shares (control) and issues certificates (economics) to the beneficiaries. So the people who receive the money are structurally distinct from the vehicle that holds the mission-lock and the §9 body's authority. The control vehicle stays disinterested; the economic certificates can be shared widely without handing anyone the kill-switch.

Proposed §9 amendment (not ratified): replace "an independent stichting that owns the operating entity" with this commercial-BV + STAK structure. The guarantees §9 requires — non-dissolvable, non-defundable, binding findings — are carried by the STAK's control lock and the equity funding, and must be confirmed enforceable by Dutch counsel (../TASKS.md).

3. What revenue is even permissible

Before how the money is split, what kinds of money are allowed — because the floors bind the business model as hard as they bind the product. This is the part a plan in this domain usually skips, and it is where the constitution does real work:

• No engagement/attention revenue. Ad-funded or time-on-site-funded models require manufacturing compulsion and surveilling users — straight through the Autonomy floor and the incident §3 honeypot discipline. Prohibited as a revenue model, not merely discouraged.
• No engineered lock-in. Revenue that depends on making exit cost more than the user was told violates the Exit floor. Subscriptions are permissible; hostage-taking is not.
• No manufactured-consent monetisation. Revenue extracted through comprehension asymmetry (confusing tiers, dark-pattern upsells) violates the Consent floor.
• The community is non-commercial by design. It is the civic commons, funded by the commercial efforts, never a product that sells access to itself. Its only permissible relationship to revenue is as a cost the commercial side carries.

So the permissible space is revenue that survives the kind/degree test: paid for plainly, exit honest, attention unmanipulated. The specific products inside that space are §8's open decision.

4. The money architecture

   revenue (permissible models, §3)
     └─ minus operating costs — including the founder's salary (a living, taken as a
        first-class cost) and the oversight costs of §5
         └─ = the profit stream
             └─ divided into EQUAL shares (certificaten) — one per party
                 ├─ each backbone organisation : one equal share
                 └─ the founder                : one equal share, the same as everyone else

• Equal shares; the founder is one party among them. The profit stream is split evenly across the equity parties. The founder holds a single equal share — there is no founder-heavy position to "dilute down from." The even split is the starting state, not a destination reached by giving a majority away. The stated aim is a living, not a fortune.
• A living is a salary, not a gamble. The founder's living is taken as a salary — a defined, first-class operating cost paid before profit is computed — not as a hoped-for dividend. The equal profit-share is upside on the same terms as every other party. This is what secures "a living, not riches" structurally: the secure part doesn't depend on a large stake, and the upside never accrues disproportionately to the founder.
• Join = re-divide evenly. When a new qualifying organisation joins, the shares re-divide so every party — the founder and every existing org alike — holds 1/N. Everyone dilutes equally, no one is protected at another's expense, and there is no reserved pool to hold back — which is exactly what the even split buys: no withheld profit waiting to be allocated, no privileged starting position to defend. The founder dilutes on identical terms to the orgs.
• Who is a party — and who is not. The equity parties are the founder and the constituency / affected-persons organisations. The financially-disinterested grader and the statutory channel hold no equity — they are paid as costs (§5). "Evenly across the parties" means evenly across the equity parties, never the oversight ones: equity in the hand that holds the kill-switch is the capture §9 forbids.
• Minting is rule-bound, not discretionary. Any organisation meeting the seat criteria joins the even split by rule; the founder cannot choose when or whether to admit a party, because that choice would be a dilution-and-capture lever.

Share conditions — the two triggers, kept distinct

The hard-won distinction (it is easy to collapse these and get capture):

• Protected against revocation-as-punishment. The entity can never strip an organisation's share because of a finding — not for ruling against it, not for voting to kill a project. Punishment is off the table; this is the §9 capture lever, denied. Oversight is exercised from the seat, without leaving, so doing the job never triggers any forfeiture.
• Forfeited on voluntary exit. Leave the coalition and the forward stream stops; the certificate is cancelled. (Distributions already received are kept — no clawback of the past, only the end of the future entitlement.) This closes the join-and-leave rent-extraction hole: a share is held while serving, not granted forever.
• Involuntary removal is rule-bound only. The line between "the org leaves" (clean forfeiture, its choice) and "the org is removed" must never be the entity's discretionary call — only a defined rule (the org dissolves, or misses defined participation criteria). Else "you're expelled" is just "your share is revoked" in a different hat.

(Residual, noted honestly: forfeit-on-leaving puts a mild thumb against resigning in protest. It is weak, distinct from in-seat oversight, and the statutory channel — Huis voor Klokkenluiders — survives any exit, so a departing org still has a non-suppressible route.)

5. Money and the kill-switch in different hands

The profit-share creates a capture risk that is the mirror of the one §9 was built to stop. §9 guards against the entity capturing the body by defunding it; sharing profit introduces capture by upside — a body that profits may not want to kill the platform, and §4 is absolute that the research must be permitted to kill it. The structure denies this the same way it denies the defunding lever — by separating the powers:

• The constituency organisations hold an equal economic share and power over the research questions (§4). Their charters run to sex-worker welfare, not to profit — partial protection, not sufficient alone.
• The grading and kill authority stays with the financially-disinterested expert and the statutory channel, who hold no equity. The upside cannot reach the people who can end a project.
• The economic certificate is framed as the population's structural stake, decoupled from oversight — held even if the org did no grading at all. So killing a project is never the killer's own financial self-harm; the disinterested party holds the trigger, and the constituency orgs' money is not contingent on the verdict.

This is the line the whole plan turns on: equal shares of the money for the founder and the population's organisations; none of the money for the hand on the kill-switch.

6. Incentive divergence, at the level of the business (core §5 / §6)

Here is where what is good for us stops being good for the user/population: the entity benefits from more revenue and more growth, and growth in this domain is historically bought with compulsion, surveillance, and the quiet instrumenting of a community as a sales funnel.

Here is what we do about it — as structure, not disclosure: the permissible-revenue floors (§3) remove the compulsion/surveillance models entirely; the community is non-commercial and not instrumented for funnel metrics; the profit stream is shared with the population's own organisations, so the entity's financial interest is coupled to, not set against, their welfare; and the people who can stop the business hold no part of its upside.

A business plan in this domain that claimed no divergence would be lying. This one locates it and answers it with where the money sits.

7. Costs, and honest sustainability

The oversight is not free, and a plan that treated it as a leftover would be the word without the thing. Funded as first-class operating cost, before profit is even computed (§4 ordering made real): the disinterested expert's grading, the channel's operation, Dutch counsel and the legal lock, the community seat's sortition process, and the community commons itself.

The founder does not hold a controlling or outsized stake. The aim, stated plainly, is a living — not wealth. That living is taken as a salary (a first-class cost, §4); the founder's equity is a single equal share that dilutes on every join exactly as the organisations' shares do. A plan in which the founder accreted the majority of the upside would be the extraction the constitution exists to refuse. This one makes the founder a partner among equals, and says so where it can be checked.

8. The open decisions

Honestly marked, because they are real and not yet made:

• The commercial products / revenue lines. What, specifically, earns the profit stream — inside the permissible space of §3. This is the central open business decision; the community (the first project) is non-commercial, so the revenue comes from commercial projects not yet defined.
• Counsel confirmation. That a STAK holding voting shares + issuing profit-certificates to an oversight coalition, with rule-based join-time minting and punishment-proof-but- membership-conditional terms, is enforceable Dutch law — and that charter-binding findings are enforceable against the founder. (../TASKS.md: "confirm the legal vehicle.")
• The share math. Who counts as an equity party (the founder + which organisations); the founder's salary level — the "living"; and confirmation that an even per-party split, with re-division on every join, is the intended shape.
• The orgs' acceptance and terms. Whether the backbone organisations will take an equity stake at all, and on what independence/conflict terms — itself a §4-governed conversation they hold power over, not an offer made to them as fait accompli.

9. Relationship to the constitution

This plan inherits and must satisfy naughtybits-core.md. It proposes exactly one amendment — the §9 legal vehicle (§2 above), in the ordinary/amendable layer — and is built to preserve every entrenched guarantee. It changes how the lock is built; it may not weaken what the lock guarantees. Until the core is amended and this plan ratified, it is a proposal: the considered commercial structure, offered for the same grading everything else here must pass.