Framework · Influence & change · 3 min read
Disagree and commit
Raise disagreement early and clearly, make the risk explicit, and once the decision is made, support it fully and agree how you will learn whether it was right.
- 1Understand
Learn the business reason. It may be legitimate: regulation, cost, deadline.
"Help me understand the goal." - 2Name risk
Make the user risk explicit, ideally with numbers. Criticise the option, not the person.
"I want to flag a concern." - 3Mitigate
Offer a phased approach, an MVP, an undo, or measuring after launch.
"Could we keep an undo instead?"
- 4Escalate?
Only for high user or compliance risk, and transparently.
"I'll take this up, and here's why." - 5Commit
Support the decision fully. Agree how you'll know if it was right.
"Our job now is to make it work." - 6Document
Log the decision, the reason, the risk and the measure.
"Review date: four weeks from launch."
Use it when
- The business decides to cut a step you think users need, for example a confirmation on bulk cancel, to hit a deadline.
- You lost a debate about a design direction and need to lead your team in the chosen direction anyway.
- A decision keeps getting reopened in hallways because people never really committed to it.
How to use it
- Understand the business reason first. Ask why the decision is being considered. It may be legitimate: regulation, cost or a deadline. Restate it so they know you heard it. Output: the reason, in their words.
- Make the risk explicit. Say what could go wrong for users, ideally with numbers: errors, task time, support load. Separate the idea from the person: criticise the option, not whoever proposed it. Output: a clear risk statement.
- Propose mitigations. Offer a way to reduce the risk without blocking the goal: a phased approach, an MVP with a follow-up, an undo, or measuring after launch. Output: one or two alternatives that keep their goal.
- Escalate only if the risk is high. If there is serious user or compliance risk and you are overruled, escalate transparently: tell the decision-maker you are doing it and why. Otherwise, let the decision stand. Output: a conscious choice to escalate or to commit.
- Commit fully and visibly. Once decided, support the decision in front of your team and stakeholders. Don't say "I told them it was a bad idea". Agree how you will know if it was right. Output: a success measure and a date to check it.
- Document the decision. Write it in a decision log: what was decided, why, the risk you raised and the measure. This makes learning possible later, without blame. Output: a decision-log entry.
Worked example
An internal operations platform. The product manager wants to remove the confirmation step on bulk cancel to make the flow faster for power users.
| Step | What the UX lead did |
|---|---|
| Understand | "Help me understand the goal." Power users cancel many orders a day and say the dialog slows them down. |
| Risk | "I want to flag a concern before we commit. With many bulk actions a day, even 1% accidental cancellations is a lot of rework for the operations team." |
| Mitigate | "Could we replace the confirmation with an undo? That keeps it fast and safe." |
| Escalate? | The PM chose to ship without the dialog first and add undo next sprint. No compliance risk, so the lead did not escalate. |
| Commit | The lead told the team: "We're shipping without the dialog. Our job is to make it work. We'll track accidental cancellations for four weeks." |
| Document | Decision log: decision, reason, risk raised, measure (cancellations reverted by support), review date. |
After four weeks the numbers were low, and the team moved on. If they had been high, the log would have made the next decision quick and blame-free.
Watch out for
- Late disagreement. Raising objections after the decision looks like resistance, even when you are right. Speak before it is decided.
- Fake commitment. Nodding in the meeting and dragging your feet afterwards destroys trust faster than open disagreement.
- Making it personal. "Your idea" versus "my idea" turns a decision into a contest. Talk about options, risks and users.
- No success measure. Without an agreed measure, nobody learns whether the decision was right, and the same argument repeats next time.
Template
Use the first part before the decision and the second part after it.
BEFORE THE DECISION
Decision being considered:
Business reason (restated):
My concern / risk (with numbers if possible):
Mitigation I propose:
Escalate? No / Yes, because (user or compliance risk):
AFTER THE DECISION (decision log)
Decided: By: Date:
Reason:
Risk raised:
How we'll know if it was right:
Review date:
- Crucial ConversationsPatterson, Grenny, McMillan & SwitzlerBook · 2002 (Amazon)
- UX Risk Statements to Influence StakeholdersNN/gVideo (external site)
"See on Amazon" links are affiliate links. As an Amazon Associate I earn from qualifying purchases, at no extra cost to you.