I sat through a standards committee where the group spent forty minutes on active listening exercises before anyone mentioned what the decision actually was. Sticky notes on the wall, a structured agenda on the screen. I have spent nearly fifty years watching variations of that scene play out under the banner of collaborative decision making, and it always ends the same way: the participants leave feeling respected, and the assumptions the decision rests on remain invisible.
Collaborative decision making is a structured conversation in which a Decider and selected participants surface and test the assumptions a decision rests on.
The goal is sufficient certainty: enough confidence in those assumptions to act, with monitoring designed before the group leaves the room. The facilitators and consultants who run these sessions earn their living from process choreography, not from examining content. If your consultancy earns its fees from two-day facilitation workshops, you will sell facilitation. Content examination requires subject-matter knowledge that most facilitators do not have.
What collaborative decision making usually means
Search the phrase and you will get technique guides and decision matrices, each with its certifications and its consulting cottage industry. What they share is a fixation on decision-making frameworks for managing the room, not for examining the decision.
I do not dismiss all of them outright. De Bono's hats do something useful when they separate creative thinking from critical thinking. None of these techniques asks the group to identify what the decision actually depends on. People leave the room feeling heard, which is pleasant, and carrying untested assumptions, which is dangerous. The facilitator writes up a summary that reads like progress. The assumptions stay exactly where they started.
Volkswagen's board approved a diesel strategy that assumed regulatory testing would only ever happen in laboratory conditions. Nobody in the governance chain surfaced that assumption for examination. When regulators tested cars on real roads in 2015, the entire strategy collapsed, cost the company more than thirty billion euros, and sent executives to prison. But you do not need a corporate scandal to see the pattern. I have watched boardrooms approve capital expenditures with polished presentations and detailed spreadsheets, where the critical assumption sat in cell B7 of a buried tab and nobody in the room thought to ask what it rested on. The spreadsheet did not lie. The assumption inside it was never examined.
One person owns the decision
A camel is a horse designed by committee. When a group shares accountability, nobody is accountable. The decision becomes an averaging exercise, and averaged decisions are mediocre by definition.
I use the term Decider for the person who owns the outcome. The Decider is not necessarily the most senior person but the person with authority and accountability for the decision's consequences. In a group setting, the Decider frames the conversation, selects the participants, and owns the result. The group surfaces and tests assumptions. They do not share the blame if things go badly. That distinction changes what the group does in the room: they stop performing agreement and start doing useful work. Useful work means naming the assumptions the decision rests on and telling the Decider when one does not hold up. That is harder than voting, and it is the only part that determines whether the decision works.
This is where collaborative decision making parts company with committee governance. Committees exist to distribute accountability. That suits board secretariats and governance teams, whose furniture and meeting cycles depend on the committee model surviving. It does not suit decision making for leaders who need a clear answer before the window closes.
I saw the result in a Dutch study on group leadership: junior-led groups outperformed senior-led groups. The finding is counterintuitive only if you assume seniority equals competence. What seniority actually provides is positional authority, and positional authority suppresses challenge. Junior leaders lack the rank to shut down dissent. If your participant list is ordered by seniority, you have built a room designed to suppress the one thing the group needs to do.
Roger Estall and I built participant selection into the Universal Decision-Making Method for this reason. Select for understanding of the subject and for the capacity to challenge group wisdom. If they will not be involved in implementation, they do not belong. Do not select for rank. Rank gets you compliance, and compliance is not what you need when you are trying to find the weak assumption that will sink the decision.
The damage runs deeper than most boards admit. I wrote separately about how rank and crowd size quietly wreck effective decision making in teams long before the first agenda item is reached.
In practice, this means the person running the purchasing system belongs in the room for a procurement decision. The non-executive who has never been inside a warehouse does not, regardless of title. This will irritate governance professionals, and I am comfortable with that.
Five steps for a group decision
The Universal Decision-Making Method gives a group five steps. Each applies whether the group has thirty minutes or three days.
- Frame the decision. The Decider states the organisation's Purpose, the scope of the decision, the authority they hold, and the constraints. Every participant enters with the same understanding of what is being decided and why.
- Develop options. The group generates possible courses of action. Not one proposal to approve or reject, but genuine alternatives that can be compared.
- Recognise assumptions. This is the step that separates the method from everything else. The group identifies what each option assumes to be true, then classifies each assumption by its influence on the outcome and the group's confidence in it. High influence, low confidence: that is the critical assumption the group must examine.
- Sufficient certainty. The group asks: do we know enough to commit? Not "are we certain?" Certainty is not available. Sufficient certainty is the point at which you know enough to act, given the stakes. If a critical assumption remains untested, the answer is no.
- Design monitoring. Before the decision leaves the room, the group agrees what to watch. Implementation proceeds as assumed. Context changes are detected. The assumptions that carried the decision are checked against reality as it unfolds.
I have compared these five steps with RACI and Delphi in a separate piece on the collaborative decision making model worth using. Role charts and consensus rounds organise participation. They do not produce a decision.
That is the process Roger Estall and I set out in Deciding. The third step, recognising assumptions, is where most group methods have a hole. They jump from options to evaluation without ever naming what the evaluation rests on. I have run this process in boardrooms and on construction sites. The scale changes. The sequence does not.
The conversation that determines the decision
Process is not nothing. A badly run conversation will produce a bad decision even with good content. But the techniques that matter are not the ones that fill facilitation manuals.
Two recur in every productive group conversation I have been part of. The first is de-biasing: active countermeasures against groupthink and confirmation bias, the two that most reliably kill honest examination. The second is the "best idea wins" ethic, which means the group's job is to find the strongest option, not to make everyone comfortable. When a group operates under "best idea wins," dissent is structural, not disruptive. The facilitation industry's version treats dissent as a problem to contain. That is precisely backwards.
Groups that create space for thinking aloud, where someone can say "I am not sure about this, but..." without penalty, surface insights that polished presentations bury. Most boardrooms have crushed that instinct.
Picture something smaller. An emergency water-pipe repair crew. The supervisor says the crew may still be on site after dark, school traffic is about to build, and asks for ideas. That took ten seconds. It gave Purpose and Context. The crew can think. A different supervisor asks where the traffic barriers should go. That got compliance, not thought. The decision fatigue that builds in organisations is often the residue of hundreds of conversations run the second way.
Consensus is when conflicting views have been rigorously tested and the best idea is self-evident. Compromise is settling for a worse idea to avoid hurt feelings. Most organisations practise compromise and call it consensus.
| Approach | Who decides | What the group examines | Typical failure |
|---|---|---|---|
| Collaborative (structured) | One Decider; group surfaces assumptions | What the decision rests on | Rare: the industry defaults to process |
| Consensus | The group collectively | Whether everyone can agree | Compromise dressed as agreement |
| Majority vote | The largest faction | Which option has more supporters | Untested assumptions in the winning option |
Most group decisions default to the third row. A show of hands, a straw poll, a reading of the room. Nobody asks what the winning option assumes to be true. Majority vote is the path of least resistance, and it produces decisions that feel democratic and rest on untested foundations.
Groups extract genuine value from partial agreement. A member disagrees with the overall direction but agrees with specific elements. A competent Decider does not dismiss the dissenter or force a package deal. They take the useful piece and leave the rest. That is precision. The governing ethic is "best idea wins," and the best idea sometimes arrives in pieces from people who disagree with each other on the whole.
Groups need a licensed challenger. Without one, assumptions go unexamined because social pressure prevents anyone from saying the obvious thing. De Bono's "black hat" approximates this, but formal contrarian roles work only when they are explicit and protected. If the designated challenger gets punished for doing the job, the role collapses into theatre and the group is back to performing agreement.
What happens after the group decides
The monitoring step is the one nobody covers in the collaborative decision making literature, which is odd, because it is the step that determines whether the decision actually works. A group that disperses after choosing an option without agreeing what to watch has done most of the work and skipped the part that matters.
Before the room empties, the Decider and the group should agree on what to monitor. If the assumption was that the new supplier could deliver within six weeks, someone watches whether six weeks is holding. If the context shifts in ways that change the assumptions the decision was built on, someone flags it. The world does not hold still while you implement.
None of my client organisations reached for their risk register when they had to decide how to respond to the disruption caused by COVID-19. They did not consult their risk appetite statement. They did not open their risk matrix. They reached for their decision-making capability: who decides, what do we know, what are we assuming, what do we watch? The register was a record of discussions held at some earlier point. It had no utility for a live decision in a changed context. That should tell you something about what group decision infrastructure is actually for.
Aviation and medicine learned this independently. Cockpit crews run structured cross-checking protocols after every significant decision. Surgical teams do the same. The principle is identical: the people who made the decision are the people who monitor it, and they monitor specific assumptions, not generic outcomes.
Every group decision failure I have examined is, at its root, an assumption failure. The group either failed to surface the critical assumption, failed to test it, or failed to monitor it after deciding.
Air New Zealand Flight 901 hit Mount Erebus on 28 November 1979. The coordinates given to the flight crew had been changed by a small administrative decision that was not properly communicated. The crew flew an automated track into terrain they could not see. 257 people died. The assumption that the coordinates were correct was never surfaced because it was never treated as an assumption. It was treated as a fact. A group that had examined the decision's assumptions would have asked: are these coordinates the same ones the crew is expecting? The answer would have been no.
After the Grenfell Tower fire killed 72 people in June 2017, the subsequent inquiry documented systemic failures across building regulation, product testing, and corporate governance. The institutional response followed the familiar pattern: a new Building Safety Act, a new regulator, layers of additional compliance. More governance, not better decisions. The assumption that the cladding system met fire safety requirements had passed through multiple hands. None of them treated it as an assumption that needed testing. They treated it as a fact established by someone else in the chain.
The Financial Crisis Inquiry Commission report documented institutional groupthink on a scale that cost the global economy trillions. The models were not wrong because the mathematics was bad. They were wrong because the assumptions underneath them were never examined by any group in the chain. Groupthink, reframed properly, is the ultimate assumption-surfacing failure: the group does not examine assumptions because the social cost of questioning them exceeds the perceived risk of leaving them alone.
The Apollo 13 crew and Mission Control in April 1970 show what collaborative decision making looks like when the stakes are existential. An oxygen tank exploded 200,000 miles from Earth. Within minutes, Flight Director Gene Kranz and his team had to decide whether to attempt a direct abort or use the lunar module as a lifeboat. The NASA review board report records how that decision followed the same structure: frame the situation, surface assumptions about power reserves and trajectory, decide. Kranz asked his team for options and got honest answers about what would and would not work. One Decider, multiple contributors, critical assumptions tested under pressure. No facilitation framework required.
The National Commission on the Deepwater Horizon found that safety inspection resources had been allocated by rig count rather than by risk profile. Deepwater operations represented a small fraction of active wells but carried a disproportionate share of catastrophic risk. The assumption that all drilling operations required similar oversight had never been examined as an assumption. It was treated as administrative common sense. When the commission forced examination of that allocation, the entire inspection regime was restructured. Not malice, not stupidity: an unexamined assumption carrying a resource decision that nobody stopped to check.
Grant Purdy is the co-author, with Roger Estall, of Deciding (2020), and the architect of the Universal Decision-Making Method.
If you have a decision you are working through, the Walk can help.
Start a Walk