Disagreeing productively

Disagreement is necessary for good decisions and is frequently avoided or conducted badly. The specific failures are identifiable.

The most common failure

Arguing about conclusions without establishing whether the disagreement is about facts, about values, or about predictions.

These require completely different responses.

Factual disagreement is resolvable by evidence, and the productive move is to identify what evidence would settle it.

Predictive disagreement — about what will happen — cannot be settled now, and the productive move is to identify what each side expects to observe and when.

Values disagreement is not resolvable by evidence at all, and pretending otherwise produces an endless exchange of facts that never touches the actual difference.

A great many long arguments are values disagreements conducted as factual ones, by people who would have finished much sooner had they noticed.

The steel man

The discipline that most improves disagreement quality.

Before responding, state the other position as well as its holder would — well enough that they agree you have it right.

Two things follow. It prevents arguing against a weaker version, which is the most common source of unproductive exchange. And it occasionally reveals that the position is stronger than you assumed, which is the point.

This is slow and it eliminates the category of argument in which both parties disagree with positions neither holds.

Identifying the crux

The specific belief on which the disagreement turns.

Two people disagreeing about a decision usually agree about most of the relevant considerations and differ on one or two.

Finding those — asking what would change your mind, and what would change theirs — converts a general disagreement into a specific question, which is frequently answerable.

Where neither side can name anything that would change their mind, that is worth knowing too, because it means the disagreement is not going to be resolved by discussion.

Why teams avoid it

The organisational problem underneath.

Research on psychological safety finds that willingness to voice dissent depends on whether people believe it is safe to do so.

Where it is not, disagreement is expressed privately after decisions rather than publicly before them, which produces both worse decisions and worse execution.

Related work on group decision-making documents pressures toward premature consensus, particularly under time pressure, with cohesive groups, and with a leader who states a preference early.

Which produces two practical rules: the most senior person should speak last, and disagreement should be explicitly invited rather than merely permitted.

Neither happens by default, and both change outcomes measurably.

Techniques that surface disagreement

Written positions before discussion. Collecting views independently prevents anchoring on whoever speaks first and preserves the diversity of judgement that is the point of consulting several people.

Assigning someone to argue against. Formalising dissent removes the social cost of it. Evidence on devil's advocacy is mixed and the practice of making disagreement a role rather than a stance is reasonably supported.

A pre-mortem. Imagining that the decision has failed and asking why. This is more effective than asking for concerns, because it grants permission to be negative.

Asking the quietest person directly. Simple, effective, and it requires the person chairing to notice who has not spoken.

Disagreeing and committing

The convention worth adopting explicitly.

Once a decision is made, the people who disagreed execute it fully, and the disagreement is recorded rather than relitigated.

This works only if disagreement was genuinely heard beforehand. Where it was not, the phrase becomes a way of suppressing dissent, which is how it is frequently used.

Recording the disagreement matters, because if the decision goes badly the record is what allows the organisation to learn rather than to reconstruct.

Personal conduct

Attack the position and not the person, which is stated everywhere and violated constantly.

Distinguish "that assumption seems wrong" from "you always assume this," which are different acts with different consequences.

And be willing to be persuaded visibly. A person who has never publicly changed their mind teaches everyone that changing your mind is a loss, which is the single most damaging norm a team can have.