Describe a time you made a decision based on data that turned out to be wrong. What did you learn?

How to answer this behavioral interview question: a complete STAR example, the mistakes interviewers watch for, and the follow-ups you should expect.

A strong STAR example answer

As a product analyst, I recommended sunsetting a feature that our usage data said only 4 percent of accounts touched. The data was clean, the dashboard was right, and leadership approved the deprecation. Within a month, we lost two enterprise clients worth a combined $250,000 in annual revenue, and three more threatened to leave.

What the aggregate data hid was concentration: the feature was used by fewer than 100 accounts, but those accounts were our largest customers, and they used it during their quarterly reporting cycle — which fell outside the 90-day window I had analyzed. My segmentation had treated all accounts equally.

Once we understood the pattern, I presented the corrected analysis, we reversed the decision, and I personally briefed the affected account teams on the timeline for restoring full support. We retained all but one of the at-risk accounts.

Before any recommendation now, I ask who the metric averages over, segment by revenue tier, and check usage against seasonal cycles, and I pair every quantitative finding with at least five customer conversations. The data told me what was happening; I had failed to ask to whom.

How to structure your own answer

• Situation: Describe the decision, the data behind it, and why it looked sound.
• Task: State your role in making or recommending the call.
• Action: Explain how the flaw surfaced and what you did to correct course.
• Result: Share the business outcome and the analytical habit you changed.

Common mistakes to avoid

• Blaming the data quality instead of your own framing and segmentation choices.
• Picking an example with no real consequence, which dodges the accountability question.
• Omitting the correction step — how you reversed or repaired the decision matters most.
• Concluding with "data can lie" rather than a concrete change in your method.

Follow-up questions to prepare for

Interviewers use follow-ups to verify your story is real. If you use this question in a real interview, expect probes like:

• How do you now validate a metric before making a high-stakes recommendation?
• Tell me about a time the data and customer feedback disagreed — which did you trust?
• How did leadership react when you brought them the corrected analysis?

Related behavioral questions

Keep drilling with these, or browse the complete question bank.