How to Answer "Describe a Data Analysis That Changed a Business Decision"

This question is the analytics professional's equivalent of a greatest hits story. Interviewers want to see the full arc: identifying the right question, selecting the appropriate methodology, executing the analysis rigorously, and—critically—communicating findings in a way that actually influenced decision-makers.

The best answers show that your analysis didn't just generate an insight—it changed behavior and created measurable value.

What Interviewers Are Really Assessing

• Problem framing: Did you ask the right analytical question, not just produce the right answer?
• Methodological rigor: Was your approach sound, or could alternative methods undermine your conclusion?
• Influence and communication: Could you translate technical findings into business language?
• Impact orientation: Did the analysis lead to action, not just a report?
• Intellectual honesty: Did you present limitations alongside findings?

How to Structure Your Answer

Follow this arc: (1) the business decision that was on the table, (2) the analytical question you formulated, (3) your methodology and key findings, (4) how you communicated the insight to decision-makers, and (5) the changed decision and its measurable outcome.

Sample Answers by Career Level

Entry-Level Example

Situation: Junior analyst whose cohort analysis changed a marketing budget allocation. Answer: "Marketing was planning to increase spend on paid social by 50% based on top-line conversion numbers showing strong performance. I was asked to build a channel attribution report and noticed something odd: paid social had high initial conversion but terrible 90-day retention—only 12% versus 45% for organic search customers. I ran a cohort analysis comparing customer lifetime value by acquisition channel. Paid social customers had an average LTV of $85 versus $310 for organic search. When I factored in acquisition cost, paid social was generating negative ROI on a lifetime basis despite looking profitable on a first-purchase basis. I presented this to the marketing director using a simple visualization: a bar chart showing cost per acquired dollar of LTV by channel. She immediately redirected $120K of the planned social increase into SEO and content marketing. Over the next two quarters, overall customer LTV improved 18% because the channel mix shifted toward higher-quality acquisitions."

Mid-Career Example

Situation: Senior analyst whose predictive model prevented a costly expansion decision. Answer: "Our company was planning to open three new retail locations based on demographic analysis showing target customer density in those zip codes. I built a predictive model incorporating not just demographics but also competitive density, drive-time analysis, and cannibalization effects from existing stores. The model predicted that two of the three locations would cannibalize existing store revenue by 15-25%, making their standalone profitability projections misleading. I presented this to the VP of Strategy using a map visualization showing overlapping trade areas and projected net revenue versus gross revenue for each location. I recommended proceeding with only one location and redirecting the capital from the other two into enhancing existing stores in underserved areas. The VP was initially skeptical because the expansion plan had board approval, but the cannibalization data was compelling enough to trigger a review. They opened one location (which performed above projections) and invested in two existing store renovations instead. The combined ROI exceeded the original three-store plan by 40%."

Senior-Level Example

Situation: Analytics director whose analysis shifted the company's entire pricing strategy. Answer: "Leadership planned a uniform 8% price increase across all product lines to offset rising costs. I conducted a price elasticity analysis using three years of transaction data, segmented by product category, customer segment, and purchase frequency. The analysis revealed dramatically different elasticities: our commodity products had high price sensitivity (elasticity of -2.1) while our premium products were nearly inelastic (-0.3). A uniform increase would drive significant volume loss in commodities while leaving money on the table in premium. I presented a differentiated pricing strategy: hold prices on commodities, increase premium products by 12%, and introduce a mid-tier option for price-sensitive customers trading up from commodities. I modeled three scenarios with confidence intervals and presented them alongside a risk analysis of the uniform approach. The CFO adopted the differentiated strategy. Over 12 months, revenue increased 11% versus the projected 6% from the uniform approach, and margin improved because the product mix shifted toward higher-margin items. This analysis established a quarterly pricing review cadence that I continue to lead."

Common Mistakes to Avoid

• Describing analysis without a decision: If your analysis produced an interesting finding but didn't change anything, it's not the right example for this question.
• Skipping the communication challenge: Generating an insight is half the job. Show how you made decision-makers understand and act on your findings.
• No counterfactual: Explain what would have happened without your analysis to highlight the value you created.

Practice This Question

Ready to practice your answer with real-time AI feedback? Try Revarta's interview practice to get personalized coaching on your delivery, structure, and content.