Tell me about persuading someone senior when the data did not clearly support either side.

How to answer

When the data doesn’t settle it, arguing harder about the data loses. The move this question rewards is changing the argument: from who is right to what it costs to be wrong, and how fast you can find out.

1. The decision and the person. Who, how senior, what they wanted and what you wanted. State both positions fairly.
2. Why the data was split. What you had and why it pointed both ways: too few cases, a hand-picked sample, a proxy metric. Show you tried to settle it with evidence first.
3. The reframe. Which option is easier to undo, what each costs if it’s wrong, and who pays. Say it in their terms, not yours.
4. The experiment. The smallest test that produces the missing data, inside their deadline, with a decision rule agreed before it runs. Agree what a mixed result means too: which part moves and which waits.
5. The outcome. What the test showed, where each of you was right, and what they did next.

Quote the decision rule as you said it: “If the new stores do no worse than the control stores, we go wide and I stop pushing back.” That line shows you were persuading toward a method, not a conclusion.

Three traps. Presenting weak data as strong to win costs you the next argument. A story where they simply gave in shows persistence, not judgment. And going over their head wins the decision and loses the relationship. The strongest version has a test that could have proved you wrong, and you say what you would have done if it had.