Running a Decision Meeting Where People Actually Disagree

Why this matters

Most shop meetings about a real decision go one of two ways. Either the owner floats an idea and everyone nods, or two strong personalities argue past each other for an hour and the meeting ends with "let's think about it." Both waste the one thing a decision meeting is for: getting the views of the people who will live with the result, in a form you can actually decide from. This article gives a six-step running order for a meeting where disagreement is expected and useful, and shows it on one real-shaped decision with the record it produces. It assumes three to eight people and a decision that matters enough to spend an hour on: dropping a service, changing on-call, adopting a price book, restructuring crews.

Why rooms default to agreeing with the owner, and the habits that counter it, have their own card. This one is the meeting procedure itself. See related: Groupthink in Partnerships and Management Meetings.

Before the meeting: send the question, not the answer

Send the question in writing at least two working days ahead, with whatever facts everyone needs. Do not send a recommendation. A memo that ends "I'm leaning toward dropping it" has run the meeting before anyone arrives.

Include the numbers people would otherwise argue about from memory: call counts, how many weekends each tech was on call last quarter, what share of calls come from plan members. A meeting that spends its first twenty minutes establishing facts has twenty fewer minutes for judgment.

Step 1: Frame the question as a choice with a decider and a deadline

Open by reading out one sentence: what is being decided, who makes the final call, and by when it takes effect. "Should we change how we cover after-hours emergency calls? The owner decides at the end of this meeting. Any change starts the first of next month."

Why: half of all meeting disagreement is people answering different questions. One person is arguing about the on-call schedule, another about whether emergency work is profitable, a third about a customer who complained last winter. Naming the question pins everyone to the same one. Naming the decider tells people their job is to inform the call, not to win a vote, which lowers the temperature and raises honesty.

What breaks if skipped: the meeting ends without a decision, because nobody knew whether it was a vote, a consensus or the owner's call. Who owns which kinds of decisions is worth settling once for the whole shop. See related: Decision Rights: Who in the Shop Decides What.

Step 2: Collect independent written views before any discussion

Give everyone three minutes to write their position, how confident they are, and the one fact they think the room is underweighting. Collect them blind and read them out together.

Why: the first spoken opinion anchors everything after it. Writing first gets each person's own view on record before anyone hears the owner's.

What breaks if skipped: you get the room's reaction to whoever spoke first, and lose the lead tech who knows something but will not say it against three people already agreeing.

Step 3: Agree the criteria before anyone argues for an option

Before options are debated, list what a good answer has to do, and get rough agreement on which matter most. Write them where everyone can see. Three to five criteria is the workable range; more than five and every option scores well on something.

Why this order: if options come first, people back an option and then pick the criteria that favor it. The tech who hates being on call will discover that "crew retention" is the most important criterion; the owner who likes the emergency revenue will discover it is "customer loyalty." Agreeing the yardstick first means the argument that follows is about how each option measures, which can be settled, not about what matters, which cannot.

What breaks if skipped: the meeting becomes a contest of values dressed up as a debate about facts, and nobody changes their mind because nobody is measuring with the same ruler. If the criteria need weighting to break a close call, that is a separate tool. See related: Scoring Options with a Weighted Decision Matrix.

Step 4: Go round the table, one point per turn, junior first

Discuss options in strict rotation: each person makes one point, then the next person speaks. Start with the most junior person and end with the decider. Keep going around until a full round adds nothing new.

Why rotation: in free discussion the two most confident people take most of the airtime, and the quiet ones never get a turn that does not require interrupting someone senior. One point per turn stops a speaker from delivering a ten-minute case that nobody can respond to in pieces.

Why junior first: the newest or least senior person has the most to lose by contradicting someone above them. Letting them speak before the senior people have committed is what makes their view reachable at all.

What breaks if skipped: the decider hears the views of the two loudest people twice and the views of the other three never.

Step 5: The decider decides, out loud, against the criteria

When a round adds nothing new, the decider states the decision and explains it against the criteria agreed in step 3: which option, which criteria drove it, and which criterion it scores worst on. Then confirm what happens next, who does it, and by when.

Why against the criteria: it shows people their input was weighed, including where it lost. "We're going with members-only because it fixes the burnout problem and keeps the customers we most want to keep, and it's weakest on emergency revenue" is a decision people can live with even when they argued against it. "Let's go with members-only" is one they relitigate next month.

What breaks if skipped: the decision gets made later, privately, and the meeting becomes a consultation people learn to stop investing in.

Step 6: Record the dissent and the trigger that reopens it

Write down who disagreed, in one sentence each, and attach a specific trigger that would reopen the decision: a number, a date or an event. Read it back so the dissenter agrees it fairly states their view.

Why: dissent that is recorded does not need to be repeated in every hallway conversation for the next six months. The dissenter knows their concern is on paper with a trigger attached, so if they turn out to be right, the shop will notice without them having to say "I told you so." And the record keeps the decider honest, because it is hard to explain away a trigger that was written down in advance.

What breaks if skipped: the losing side keeps the argument alive informally, and when the decision starts to wobble nobody can tell whether it is failing or whether the people who opposed it are just louder.

Worked example: after-hours emergency calls

A five-person management group at an HVAC and plumbing shop: the owner (decider), the ops manager, the office manager, and two lead techs. The prep memo, sent two days ahead, states that after-hours calls run about 6 a week in winter and about 3 a week in summer, that 40% of them came from maintenance plan members last year, that the four techs on the rotation each covered about one week in four, and that two of those four have said on-call is why they are looking elsewhere.

Step 1 framing: "Do we change after-hours coverage? Owner decides today. Change starts the first of next month."

Step 2 blind cards:

Five people, four distinct positions, and the owner the least sure of anyone except the ops manager. In a spoken round opened by the owner this would likely have been five people agreeing to keep it.

Step 3 criteria, agreed before options: keep the rotation techs; honor what plan members were sold; keep emergency work that leads to bigger jobs; avoid losing customers to competitors. Ranked in that order after a short exchange, because a lost tech takes longest to replace.

Step 4 rotation, junior first. The useful points that surfaced: members-only would cut after-hours volume by about 60%, since 40% of calls came from members, leaving about 2.4 calls a week in winter and 1.2 in summer; the replacement sales the ops manager cited came mostly from members anyway; and the plan agreement's priority language turned out to cover next-business-day service, not nights, so the office manager's "promise" was softer than the room thought. Lead tech A conceded members-only fixes most of the burnout problem. Lead tech B did not.

Step 5 decision: members-only after-hours coverage, rotation unchanged in shape but carrying about 40% of the old call load. Best on keeping techs and on members, adequate on replacement leads, weakest on competing with 24/7 advertisers.

Step 6 record:

DECISION   2026-02-10   owner decides
Question: change after-hours emergency coverage?
Decided: plan members only, from 03/01
Drove it: rotation burnout, member promise
Weakest on: non-member emergency demand
Owner: office mgr updates scripts and website
DISSENT
  Lead tech B: rotation still breaks sleep;
  wants nights dropped entirely.
REOPEN IF any one of:
  - average night pages exceed 3 per on-call week
    over any 6-week stretch
  - a rotation tech resigns citing on-call
  - member renewals fall 10 points vs last year
Review date: 2026-06-01

What would have changed the decision. If plan members had been only 10% of after-hours calls, members-only would barely have reduced the load and the room would have been choosing between keeping it as is and dropping it. If the plan agreement had promised night coverage in writing, members-only would have been the floor, not an option, and the question would have been how to staff it. Neither condition held, which is why the members-only option won.

How to verify the meeting worked

• The blind cards showed at least two positions. If they are unanimous every time, the cards are being written to match the owner, or the meetings are being called on questions that are already decided.
• The criteria were written before the first option was argued, and the decision statement named which criterion the chosen option was weakest on.
• Everyone spoke at least once in rotation, and the decider spoke last.
• The dissent record has a trigger a stranger could check. "If it isn't working" fails. "Average night pages over 3 per on-call week for 6 weeks", set above the roughly 2.4 a week winter now expects, passes.
• On the review date, someone pulled the trigger numbers, whether or not anyone remembered the meeting. A trigger nobody checks is a courtesy, not a record.

References

• See related: Groupthink in Partnerships and Management Meetings
• See related: Scoring Options with a Weighted Decision Matrix
• See related: Decision Rights: Who in the Shop Decides What
• See related: Keeping a Decision Journal
• See related: Writing Decision Rules So the Office Stops Asking You