Business Problem-Solving & Banking Scenarios
The skill that separates analysts from coders: turning a vague business question into metrics, hypotheses, and a recommendation a stakeholder can act on.
Learning Objectives
- Follow a repeatable business problem-solving framework.
- Distinguish a metric from a KPI.
- Diagnose why a metric changed using a structured approach.
- Apply the framework to banking scenarios (churn, sales drop, fraud).
- Turn analysis into a clear recommendation.
1. The Business Problem-Solving Framework
Every business case question follows the same five steps. Memorize them:
- Clarify — restate the business question and ask what success looks like.
- Define metrics — choose the KPIs that map to that question.
- Hypothesize — brainstorm possible causes/levers.
- Analyze & segment — test hypotheses by breaking data down.
- Recommend — state the insight and what to do.
2. Metric vs KPI
| Metric | KPI | |
|---|---|---|
| What it is | any measurement | a metric tied to a business goal |
| Example | number of logins | monthly active users (MAU) |
| All KPIs are metrics, but not all metrics are KPIs | — | — |
A KPI is a metric you've decided matters to a goal — it's what you put on a dashboard.
3. "Why Did This Metric Change?" — the Diagnosis
This is the most common business-case prompt. Work through these four buckets in order:
- Data problem? — pipeline broken, logging change, definition changed?
- Natural? — seasonality, weekdays vs weekends, cyclical patterns?
- Internal? — price change, product change, policy, marketing?
- External? — competitor move, market shift, macro event?
4. Banking Scenarios — Practice
Apply the framework to these generic banking/financial-services cases. For each, walk through the five steps.
5. Interview Questions (with Model Answers)
The business-case questions interviewers ask. Self-test before revealing.
IQ1. Walk me through how you'd approach a business problem with data.
Model answer: "I'd clarify the business question, define the right metric, brainstorm hypotheses, analyze by segmenting the data, then recommend an action. That keeps me from jumping to analysis before I know what the stakeholder needs."
IQ2. What's the difference between a metric and a KPI?
Model answer: "A metric is any measurement; a KPI is a metric tied to a business goal. Every KPI is a metric, but not every metric is a KPI — KPIs are the numbers I'd put on a dashboard."
IQ3. A key metric dropped last month — how do you diagnose it?
Model answer: "First I'd confirm it's real and not a data or definition issue. Then I'd check if it's seasonal, then whether an internal change (price, product, policy) or an external factor (competitor, market) caused it — segmenting the data to isolate the driver."
IQ4. How do you choose the right metric for a business question?
Model answer: "I map the metric to the business goal — what behavior or outcome reflects success? I'd choose a metric that's measurable, tied to the decision, and not easily gamed."
IQ5. What banking metrics would you track and why?
Model answer: "Customer metrics like churn rate and lifetime value, transaction metrics like volume and average value, and risk metrics like default or fraud rate. The exact set depends on the business goal."
IQ6. How do you structure a recommendation from an analysis?
Model answer: "I state the insight in one line, explain the evidence, then give a specific, actionable next step — with the expected impact. A recommendation should be something a stakeholder can act on."
IQ7. How do you handle a vague business request?
Model answer: "I ask clarifying questions — what decision is this for, who's the audience, what time frame and segments matter. Clarifying first prevents building the wrong analysis."
IQ8. Revenue fell month-over-month — what's your first check?
Model answer: "Whether it's fewer transactions or lower average value, and whether it's concentrated in a specific region, product, or segment. That segmentation tells me where to look next."
IQ9. In business terms, what does "correlation ≠ causation" mean?
Model answer: "Two metrics can move together without one causing the other — a third factor may drive both. So before recommending action, I'd check for confounders or run an experiment."
Hands-On Project: Solve a Business Case
Work through a banking case using the five-step framework.
The prompt
"Online account signups dropped 30% this month. What would you do?"
- Write 2 clarifying questions you'd ask.
- Name the metric you'd track.
- List 3 hypotheses.
- Describe how you'd segment the data to test them.
- Write a one-line recommendation (assuming you found a checkout bug).
View Solution / Walkthrough
- Clarify: "Which signup channel dropped? Is it a specific device or region?"
- Metric: daily new signups, segmented by channel/device/region.
- Hypotheses: (a) a checkout bug, (b) a marketing campaign ended, (c) a competitor launched.
- Segment: compare signups by channel and device; check for a sudden drop vs a gradual decline.
- Recommend: "The drop traces to a checkout error on mobile — fix it and monitor signups to confirm recovery."
Key Takeaways
Clarify → define metric → hypothesize → analyze → recommend.
A KPI is a metric tied to a goal; every KPI is a metric.
Confirm a metric change is real before explaining it.
Segment the change to isolate the driver (region/segment/product).
End with an actionable recommendation, not just a finding.
Objective Questions — Test Your Understanding
Q1. What is the FIRST step in a business case question?
Q2. What's the difference between a metric and a KPI?
Q3. When a metric drops, what should you check FIRST?
Q4. In the framework, which step comes after defining metrics?
Q5. "Sales dropped in the South" — the best first move is to…