“Tell me about a time you used data or analytics to make a decision” is testing far more than whether you can read a dashboard. The interviewer wants to understand how you turned evidence into a judgement, what alternatives you considered and whether the final decision produced a useful outcome.
Closely related questions include “How have you leveraged data to develop a strategy?”, “Tell me about a time you used data to influence a business decision” and “Give an example of when you used data analytics or technology to identify inefficiencies or drive cost improvements.” The same preparation can cover all of them.
Want more interview answers to practise with?
The full guide contains 250+ interview questions and sample answers, including common behavioural and decision-making questions.
What the interviewer is really testing
📊 Evidence selection: did you use the right data rather than whatever happened to be available?
🧠 Interpretation: did you understand what the data meant and where its limits were?
⚖️ Judgement: did you compare realistic options rather than treating the numbers as an automatic answer?
🗣️ Communication: could you explain the evidence clearly enough for other people to act on it?
📈 Outcome: did the decision improve performance, cost, service quality or another meaningful result?
A weak answer says, “I looked at the data and chose the best option.” A stronger answer explains which data mattered, what patterns you found, what other evidence you checked, which options you considered and why the chosen course was proportionate.
How to structure your answer with B-STAR
💭 Belief: briefly explain your approach to evidence-led decisions.
📍 Situation: identify the decision and why it mattered.
🎯 Task: state what you personally had to decide or recommend.
🛠️ Action: explain the data you gathered, how you analysed it, what alternatives you considered and how you handled uncertainty.
📈 Result: show the effect of the decision and how you knew it was successful.
The Action should carry most of the detail. Interviewers are normally more interested in how you used the evidence than in the technical name of the analytics tool.
Example answer: using analytics to allocate a marketing budget
Belief: I believe data is most useful when it helps compare options objectively, but I also check that the underlying measures are relevant before making a decision.
Situation: In a marketing role, I was planning a product launch with a limited budget. We could invest across social media advertising, email, content and paid search, but spreading the budget evenly would have reduced the impact of every channel.
Task: I needed to recommend how the budget should be allocated to achieve the strongest return.
Action: I reviewed historical campaign performance, including conversion rate, acquisition cost and audience engagement. I compared this with competitor activity and set up tracking so we could monitor live results after launch.
I found that two channels consistently produced stronger conversion at a lower acquisition cost, but one had a much smaller available audience. Rather than simply putting the full budget into the cheapest channel, I modelled several allocation options and considered reach, likely conversion and diminishing returns.
I recommended concentrating spend on the two strongest channels while keeping a smaller test budget for a third channel where the historic evidence was weaker but the target audience was growing. I explained the assumptions clearly and agreed review points so we could move spend if live performance differed from the forecast.
Result: The launch exceeded the original sales projection by 20%. The strongest channels delivered the majority of conversions, and the live review points allowed us to move budget away from underperforming activity quickly.
Example answer: using data to identify an inefficient process
Belief: I believe data can expose inefficiency that is hard to see from individual cases, particularly when teams have become used to working around the same problem.
Situation: I supported an operational service where case turnaround times were increasing. Staff believed the main problem was overall demand, but the backlog was growing faster than incoming volumes.
Task: I needed to identify where time was being lost and recommend a practical improvement without increasing headcount.
Action: I analysed case timestamps, rework rates and hand-off volumes across each stage of the process. The data showed that one approval step was responsible for a disproportionate amount of delay because cases were frequently returned for missing information.
I checked a sample of returned cases to understand why. Most were missing the same small group of evidence items, so I tested whether an earlier validation check would prevent rework. I compared the likely time added at the start of a case with the much larger delay created by a failed approval later.
I recommended moving the evidence check to first contact and creating a standard checklist. I presented the expected saving to managers, then monitored turnaround time and rework for four weeks after implementation.
Result: Returned cases reduced, average turnaround improved and the team cleared more work with the same staffing level. The data also gave managers a better basis for deciding where future process improvement effort should go.
Practise more common interview questions
Use the full guide to compare different ways of structuring evidence-led, behavioural and general interview answers.
How to answer if the question mentions cost improvements
If the interviewer asks how you used data analytics or technology to identify inefficiencies or drive cost improvements, avoid turning the answer into a finance exercise unless that is genuinely what happened. Cost improvement can come from reduced rework, lower manual effort, fewer errors, better use of capacity or avoiding unnecessary spend.
💷 Explain the original cost or inefficiency.
📊 Show which data proved where the problem sat.
⚙️ Explain the change you recommended.
📈 Quantify the benefit where you reasonably can.
Common mistakes
❌ Listing tools instead of judgement: naming Excel, Power BI or SQL does not explain how the decision was made.
❌ Using one metric in isolation: explain why the measure was relevant and what other evidence you checked.
❌ Confusing correlation with a decision: a pattern may suggest a cause, but show how you tested the assumption.
❌ Hiding uncertainty: stronger answers acknowledge imperfect data and explain how the risk was managed.
❌ Giving no alternatives: show what other options were considered before you chose the final approach.
❌ Ending with “the data proved me right”: explain the actual business or service outcome.
Quick interview check
✅ Is the decision clear?
✅ Have you explained why you chose those data sources?
✅ Did you interpret the evidence rather than simply quote figures?
✅ Have you shown the alternatives or trade-offs?
✅ Did your decision produce a measurable or clearly evidenced result?
Prepare for the questions employers ask most often
