Governance
and Power.

The creator tool has customers now. The independent creator community contributed ideas and guides. The publication helped people understand the release. A distribution partner introduced buyers under a promised referral split. The product team proposes a new split for the next campaign: more money for the team, less for the partner.

The proposal may be reasonable. The team now carries support costs it did not anticipate. It may also be self-serving. The partner committed time and reputation under the old terms. The community wonders whether its unpaid work helped create value that neither party’s referral calculation sees. A decision about percentages has become a decision about whose expectations count.

Power is the ability to change another participant’s possibilities. Governance makes that power visible and answerable.

Find the power before writing the rule.

A person who can set a commission policy has power. So does the person who decides which contributions enter the record, who may join a game, which agent may act for a group, and which dispute reaches a reviewer. Some power is obvious in an administrator control. Some hides in defaults, missing fields, private data, or the ability to delay a decision until the other party cannot afford to wait.

The creator community should be able to distinguish four questions. Who can propose a change? Who can make it binding? Who is affected by it? Who can inspect or challenge the result? Giving one actor the technical ability to update a rule answers only the second question. It does not explain why that actor should decide alone.

Power also follows information. If the product team sees every sale but the partner sees only a final payout, the partner cannot test the calculation. If the community cannot see how its work was represented, it cannot correct the account. Useful governance gives participants enough visibility to understand a consequential decision while protecting buyers’ and creators’ private information.

Make authority specific.

The partner may have permission to distribute the tool and earn a commission. That does not give it authority to speak for the creator community. The product team may own the product and its budget. That does not give it the right to rewrite a partner’s completed work under new terms. A reviewer may approve a publication without becoming its author. Each authority has a subject, an action, a scope, and a way to end.

A good design records the source of that authority. Was it a role assigned when someone joined? An agreement between two groups? A limited mandate to act for someone else? A decision made by a body whose members can be identified? The answer matters most when a person says no. Refusal should show which boundary held, not merely that a button failed.

For the proposed split, the team might be authorized to set terms for future campaigns. A past referral should remain governed by the terms under which it was made. If the group wants to alter an unsettled obligation, it needs the affected party’s agreement or an earlier agreed procedure. A timestamp alone does not make a change fair; the time at which people relied on the rule matters.

Change rules in the open.

Suppose the new split would begin next month. The proposal should state the current rule, the proposed rule, why it is changing, when it would take effect, which commitments it covers, and who can approve it. Participants need a period to ask questions while they can still change their plans. A game’s rules may evolve; its members should not have to discover an amendment through a smaller payment.

The right procedure depends on the stakes. A team can fix a spelling error in a help page quickly. Changing the formula for a commission, a publication gate, or access to customer relationships deserves a higher bar. Some changes may require the consent of every affected party. Others may be decided by a named steward after notice and review. A community may choose a vote, but a vote is not automatically sufficient when a minority has already contributed under a specific promise.

Keep the earlier rule and the reason for the revision. A participant should be able to ask which version applied to a particular action and why. If a decision is reversed, preserve the reversal as part of the history. An editable policy with no readable history makes even a sensible amendment look like an arbitrary command.

Give a challenge somewhere to go.

The partner disputes one referral. The buyer used its link, but the record assigned the sale to a later introduction. The narrow question is whether the agreed attribution rule was applied correctly. The broader question is whether that rule captures the kind of contribution the partners care about. Those questions need different remedies: correct a mistaken calculation now; reconsider the rule for future work.

A challenge needs a claimant, the decision being questioned, the evidence each side may inspect, a reviewer with authority to act, and a recorded outcome. The reviewer may need to pause settlement while the facts are checked. Where the first decision maker benefits from the outcome, an appeal should reach someone else. A decision that can be contested only by asking the same actor to reconsider is a weak safeguard.

There are limits to what a record can settle. A referral touch can establish that a link was used; it cannot prove that the link alone caused the purchase. A review event can establish that a draft was approved; it cannot prove that the published account was fair. Governance must allow judgment while making the evidence, authority, and limits of that judgment inspectable.

Let people leave with agency intact.

If the partner rejects the new terms, it should be able to stop making future referrals without losing payment already earned. A creator should be able to leave a shared project with clarity about their own work and the uses they already permitted. Exit may require a transition for customers, unfinished obligations, or safety. It should not become a threat that forces someone to accept any new rule.

Independent minds will sometimes disagree about purpose. Their connection is healthier when the terms of separation are as clear as the terms of joining: what stops, what continues, what records each party may keep, and what can still be challenged. The ability to leave gives consent meaning beyond the moment of entry.

One implementation, an unfinished question.

Opentangle provides concrete pieces of this anatomy. MoM maintains durable game identity, memberships, role claims, and shared state. A Game Master interprets the game’s rules. The Attention Protocol has review and publication gates. The Coseller Protocol lets the Game Master set an attribution policy, retains policy history, and records settlement actions. The Mandate Protocol makes representation explicit, bounded, and revocable.

These are mechanisms of authority and accountability, not a general constitution for every game. The current Coseller policy control gives the Game Master the decision, not a built-in vote or consent process for all affected cosellers. Its audit trail does not by itself supply an appeal, and an already settled payout has no general clawback path. Mandate governs a particular representation relationship; it does not decide how a community amends its collective rules. Other implementations may choose different institutions and records.

The portable design question is therefore open: when a group changes what others may do or receive, can it name who decided, under which authority, on which evidence, and with what route for challenge or exit?

Make the rule legible. Make the decision accountable. Let the affected minds answer back.