Skip to content

Conditions for Self-Organization to Work

Definition

Open Space Technology is not a universal meeting format — it works, and should only be used, when five "Conditions of Use" hold: a real business issue is at stake (not a manufactured exercise), the issue carries a great deal of complexity, there is real diversity among the people and viewpoints involved, real passion (and likely real conflict) is present, and there is genuine urgency — "a decision time of yesterday." Where the answer is already known, or where someone in authority needs to remain in control, Owen states plainly that Open Space "will not work, and therefore should not be used."

In the Book

Chapter II ("Preparation") opens with this diagnostic test before any logistics: Owen contrasts installing a pre-designed corporate billing system (no room for Open Space — "simply issue the manuals") against a genuinely unresolved question like "the future of the corporation," where even senior leadership admits it doesn't know the answer. Later, in the closing chapter, Owen connects this back to Stuart Kauffman's work on self-organizing chemical and biological systems — "order for free" — arguing that Open Space works for the same underlying reason those systems do: certain preconditions, when present, allow a system to spontaneously reorganize itself into a higher order of complexity without a central designer. The book's own case data (sixty thousand-plus uses across 124 countries) is offered as evidence that when the five conditions are met, the self-organizing outcome is reliable and repeatable, not a fluke.

Why It Matters

This concept guards against treating "just let it self-organize" as a universally applicable technique. It supplies a checklist for diagnosing, in advance, whether a situation actually has the raw material — genuine stakes, genuine complexity, genuine diversity of perspective, genuine motivation, genuine time pressure — that self-organizing processes need to feed on. Applied outside meetings, it is a general answer to "when does decentralizing control actually work, versus when does it just produce noise": the mechanism is not magic, it is conditional, and naming the conditions makes the failure mode diagnosable rather than mysterious.