Negotiation and Debate - Mistake¶

When to use it¶

• Use the full loop for a decision with real trade-offs, several teams, or a lasting effect. It makes needs, evidence, and ownership clear before the disagreement turns personal.
• Use a short version for a small, reversible choice. State the decision, hear the key concern, choose an owner, and move on; a full workshop would cost more than a wrong answer.
• The benefit is a decision people can explain and support. It reduces arguments based on title, volume, or who spoke first.
• The cost is time and preparation. Gathering evidence and hearing interests takes effort, so time-box the discussion and match it to the risk.

Common mistakes¶

• Arguing positions as if they are the real problem. "Friday" and "two weeks" hide the customer need and safety concern, so the discussion becomes a tug-of-war. Fix: ask what each side needs, fears, and would accept before proposing a solution.
• Making the disagreement about a person. Personal language makes people defend themselves instead of testing the idea. Fix: name the claim, evidence, and impact; do not name a person's character or motive.
• Using compromise as a shortcut. Splitting the difference can still fail the customer need or the safety limit. Fix: generate distinct options and test each one against shared criteria.
• Calling opinions evidence. Confidence, seniority, or a past anecdote can hide what has not been measured. Fix: agree on the relevant test, customer fact, cost, risk, or precedent before choosing.
• Debating without a decision owner or deadline. The same points return in the next meeting because nobody can end the discussion. Fix: state who decides, when they decide, and what information they need.
• Inviting feedback without scope. People either stay quiet or raise every possible concern, neither of which helps the current choice. Fix: state the question, what feedback is useful, and what is already decided.
• Silencing dissent to keep the meeting pleasant. The hidden concern comes back during delivery, often when changing course costs more. Fix: ask for the strongest case against the leading option, and give people quiet writing time before discussion.
• Treating a decision as permanent when the facts change. A new test result, customer constraint, or incident can invalidate the original choice. Fix: record the review trigger, then reopen only when that trigger or other material evidence appears.
• Keeping the argument alive after a fair decision. Repeating a losing argument drains trust and delays delivery. Fix: challenge respectfully before the call, then commit to execution; Amazon describes this as "disagree and commit."
• Using blame after an option fails. Blame makes people hide facts and become risk-averse. Fix: describe what happened, the conditions that allowed it, and the next system or process change, as recommended by Google SRE.

Continue to Best Practise.