Driving the Case and Asking for the Right Data
A messy case feels like standing in a control room with every dashboard flashing at once. The weak candidate asks for “more data”; the strong candidate asks for the one number that can kill or confirm the current hypothesis.
- Driving the case means owning the path from issue to recommendation, not waiting for the interviewer to rescue you.
- The best data request has three parts: hypothesis, metric, decision use.
- Ask for data only after saying why it matters: “To test whether demand or margin is the issue, may I see volume and gross margin by channel?”
- Good case leadership is a loop: frame the issue, request data, extract insight, decide the next branch.
- Never boil the ocean. Prioritise the largest profit pool, fastest-changing trend, or most decision-critical uncertainty first.
- When you receive data, pause and say the implication before asking for the next exhibit.
- The common trap is becoming a data collector instead of a problem solver.
Big Picture: You Are Not Asking for Data, You Are Steering a Decision
In a candidate-led case, data is not a shopping list. It is a steering wheel. Every number you request should reduce uncertainty, eliminate a branch, or move you closer to a recommendation. If you need a refresher on case ownership, revise Interviewer-Led versus Candidate-Led Cases before practising this skill.
Core Explanation: The Data Request Loop
The cleanest way to drive a case is to treat every move as a loop: state the issue, choose the branch, request the right data, convert it into insight, then decide what to test next.
Suppose the case says, “A food delivery platform’s profit has fallen.” A weak first request is: “Can I get revenue, costs, customers, geography and competitors?” It sounds active, but it gives the interviewer no logic.
A stronger request is: “I would first split profit decline into revenue and cost drivers. To see whether the issue is demand-side or unit economics, may I see order volume, average order value and contribution margin over the last few periods?”
Notice the difference: the second answer tells the interviewer what you are testing, which metrics matter, and what decision the data will support.
The Three-Part Formula for Asking the Right Data
Use this sentence structure until it becomes muscle memory:
“To test whether [hypothesis], I would like [specific data] so that I can decide [next action].”
The Data Menu: Six Metrics Worth Asking For
In profitability, growth, market-entry and operations cases, these are the measures candidates most often need. Do not memorise them as a list. Use them as a menu linked to the business question.
Interviewers like candidates who ask for a cut, not just a metric. “Revenue” is broad. “Revenue by product line for the last four quarters” is useful. “Gross margin by channel versus last year” is even better when you are testing whether channel mix caused the profit drop.
Definitions You Can Say in One Breath
- Driving the case: Actively guiding the problem-solving path from structure to recommendation using hypotheses, evidence and prioritised next steps.
- Right data: The minimum information needed to confirm, reject or size the most important current hypothesis.
- Data cut: A specific breakdown of a metric by time, segment, product, geography, customer type or channel.
- Insight: A business implication drawn from data that changes what you would do next.
How to Know Which Data to Ask First
When everything looks important, prioritise by decision value. Ask yourself: “If I had this number, would it change my next move?” If the answer is no, delay it.
A useful rule: data without a comparison rarely gives insight. “Margin is 18%” is a fact. “Margin fell from last year and is lower in offline channels” is a direction.
Mini Case Study: Wakefit and the Data Behind an Omnichannel Move
Wakefit shows how a digital-first Indian brand can use customer and channel data to decide where offline experience matters most.
Wakefit built its brand in a category where online discovery can work, but final purchase confidence often depends on touch, comfort and trust. For mattresses and furniture, the customer wants to feel the product, compare firmness, ask delivery questions and reduce the fear of a wrong high-value purchase.
A poor case approach would be: “Should Wakefit open more stores? Let us ask for online sales, offline sales, marketing spend and competitors.” That is too broad.
A sharper case driver would split the problem into three uncertainties: where offline improves conversion, which cities justify fixed costs, and whether offline stores support online sales rather than simply shifting revenue from one channel to another.
The primary driver here is not “offline expansion” by itself. The primary driver is using existing demand signals to place experience-led assets where they reduce customer uncertainty. Supporting drivers include category economics, city-level density, customer trust, delivery capability and the ability to connect online and offline journeys.
So what: in a case, you do not win by asking “Can I have channel data?” You win by asking for the channel data that tells you whether offline presence creates incremental profit, improves conversion, or merely adds cost.
How AI Changes Driving the Case and Asking for the Right Data
AI changes case preparation in three practical ways, but it does not remove the need for judgment.
- AI can simulate interviewer pushback. You can ask ChatGPT or Claude to interrupt you when your data request is too broad, forcing you to justify why the number matters.
- AI can generate alternative data cuts. For a profitability case, it can suggest cuts by product, region, customer cohort, channel and time period, helping you practise sharper requests.
- AI can review your transcript for case leadership. It can flag moments where you asked for data without a hypothesis, skipped interpretation, or failed to choose the next branch.
Use ChatGPT as a mock interviewer: paste a case prompt, ask it to reveal data only when your request is specific, then ask it to score every request as “decision-useful” or “too broad.” For a full practice routine, use Practising Cases With AI as a Mock Interviewer.
Interview Relevance
“Our client is a mid-sized apparel retailer. Revenue is growing, but profits are declining. What data would you ask for first, and why?”
After every exhibit, say: “This tells me…” before saying “I would next like…”. That one habit makes you sound like a consultant, not a spreadsheet tourist.
Common Mistake
The mistake: asking for every possible number to look thorough. Why it costs you: it signals weak prioritisation and forces the interviewer to lead the case for you. One-line fix: before every data request, say the hypothesis it will test and the decision it will support.