Skip to main content
Beta Library · How people work in it

Deciding without a boss

The first question anyone asks about self-organization: if nobody is in charge, who decides? The answer is not "everyone together" — that would be slower than hierarchy, not faster. It is that different kinds of decision get different methods.

The misunderstanding worth clearing up first

Decentralized decision-making is routinely confused with consensus — everyone discusses until everyone agrees. That model deserves its bad reputation: it is slow, it rewards whoever has the most stamina, and it quietly gives a veto to anyone willing to keep objecting.

None of the schools in this library propose it. What they propose is a distinction: most decisions need one decider, a few need a group, and the difference is decided by scope, not by seniority.

Not this

❌ Everyone decides everything

  • Every question goes to the whole group.
  • Agreement required before anything moves.
  • Slower than hierarchy, and exhausting.
But this

✅ The method fits the decision

  • Routine calls: whoever holds the role, alone.
  • Consequential calls: one decider, after seeking advice.
  • Rules that bind everyone: the group, by consent.

Four methods

1. Role decision

For: everything inside someone's own responsibility

Whoever holds the role decides, alone, without asking. This covers the overwhelming majority of decisions in any organization, and it is the main source of speed. If this one is not genuinely in place, none of the others help — people will keep checking upward out of habit.

2. The advice process

For: consequential decisions affecting others

One person decides — but must first seek advice from those affected and from people with relevant expertise. The advice is not binding, and there is no vote. What is binding is asking, and being able to show that you did.

This is the workhorse of most self-managed organizations, and its elegance is that it keeps decisiveness while removing the two failure modes of hierarchy: the uninformed decision from above, and the decision made in isolation.

2b. The consultative individual decision

A close relative, worth naming separately

The BetaCodex version: consultation is an obligation, but there is exactly one decider — deliberately not the group. The emphasis differs slightly from the advice process. Here the point is to prevent decisions from drifting into committees; there the point is to prevent them from being made in a vacuum.

3. Consent

For: rules and agreements that bind a whole group

A proposal is adopted when nobody has a reasoned objection — an argument showing avoidable harm or risk. Note what that is not: it is not "everyone thinks this is the best idea". Reservations are recorded but do not block. The standard is good enough for now, safe enough to try, and every decision carries a review date, so it is a bet with an expiry rather than a permanent law.

This is where the difference between an objection and a concern does the real work: without it, consent degenerates into consensus and inherits all its problems.

4. Election by consent

For: filling a role

Nominations are made openly and with reasons, people may change their nomination after hearing others, and the resulting proposal goes through the same objection round. Crucially, roles are held for a fixed term — which addresses a problem hierarchies share: the person who has held the important role for a decade and cannot be moved, because nobody ever agreed when it ends.

What a consent round actually looks like

TensionSomething isn't working. Described observably, not as a complaint.
ProposalA concrete suggestion for what should apply instead.
QuestionsClarifying only — not yet opinions.
ObjectionsReasoned harm or risk. Reservations noted separately.
AdoptedAll objections integrated. With a review date.

The step people skip is the first one. A proposal without a stated tension invites debate about preferences; a proposal that names what is not working invites debate about the problem. That single discipline changes the character of the meeting more than any facilitation technique.

SignalWhat it meansEffect
Objection"This causes concrete, avoidable harm — here is why."Blocks until integrated
Concern"I'm uneasy about this, but I can't name a specific harm."Recorded, does not block

Where it gets hard

Habit outlasts structure. The most common failure is not that decisions go wrong, but that they never get made locally. People keep asking upward, and whoever used to decide keeps answering — because both sides find it more comfortable. Removing the formal authority does not remove the reflex.

Advice can become a veto in disguise. If seeking advice turns into needing approval from everyone consulted, the process has quietly become consensus. The test is whether anyone has ever decided against the advice they received and been backed up afterwards.

Meetings can eat the gains. Holacracy's structured governance meetings are the cautionary example: enough procedure that the overhead becomes its own complaint, and one of the reasons Zappos rolled the model back. Keep the mechanism minimal.

Nothing here removes power. Even with immaculate process, some voices carry more weight — through expertise, tenure or personality. Written decisions with named participants at least make that visible, which is more than an informal culture offers.

The point is not that everyone decides everything. It is that the decision sits where the knowledge is — and that the method is chosen by what is at stake, not by who has the title.

How BetaOS relates to this

Roles are in the product, so role decisions have a visible home: who holds which responsibility is documented rather than assumed. The governance view records how a cell handles agreements, participation and conflict — as an observation across cells, without a score.

What does not exist yet is a place to record decisions themselves — tension, proposal, objections, review date, and the successive versions of an agreement. It is designed and deliberately deferred: a decision log is exactly the kind of artefact that gets filled in enthusiastically for three weeks and then abandoned, and the honest precondition is that the simpler governance view actually gets used first.

Sources

  • Bernhard Bockelbrink, James Priest, Liliana David: Sociocracy 3.0 — A Practical Guide. CC BY-SA — consent, objections, role selection.
  • Frederic Laloux: Reinventing Organizations. 2014 — the advice process, drawn from AES and Morning Star.
  • Niels Pflaeging, Silke Hermann: Cell Structure Design — the consultative individual decision.
  • Morning Star Self-Management Institute — peer agreements and the stepwise conflict process.
  • Brian Robertson: Holacracy Constitution v5.0 — integrative decision-making, and the cost of procedure.

More from the Beta Library: betaos.org/library