Business Case Studies — Interview Preparation

Case study questions test whether you can take a vague business problem, structure your thinking, identify the right data, and communicate a clear answer. These are common in data analyst and product analytics interviews.

Tags: #BusinessAnalysis #CaseStudy #InterviewPrep #Analytics

---

How to Approach Business Cases

A good case answer has four parts:

1. Clarify — ask questions before diving in. What does "improve" mean? What's the time frame? What data do we have?
2. Structure — break the problem into parts. What are the possible causes? What are the key metrics?
3. Analyse — hypothesise and prioritise. Which cause is most likely? What would you check first?
4. Recommend — give a concrete next step with a reason.

---

Case 1 — Revenue Decline

"Our monthly revenue declined 12% last month. Investigate."

Show answer

Step 1 — Clarify: - Is this a one-time drop or a continuing trend? - Compared to what baseline — last month, same month last year, forecast? - Has anything changed recently (product, pricing, marketing, competitors)?

Step 2 — Segment the decline:

Dimension	Questions
Time	Did it drop suddenly (event) or gradually (trend)?
Geography	Is it global or region-specific?
Product	Across all products or one category?
Customer	New vs returning? A specific segment?
Channel	Web, mobile, marketplace, in-store?

Step 3 — SQL skeleton:

-- Compare this month vs last month by category and channel
SELECT
    category,
    channel,
    SUM(CASE WHEN month = '2024-02' THEN revenue END) AS this_month,
    SUM(CASE WHEN month = '2024-01' THEN revenue END) AS last_month,
    SUM(CASE WHEN month = '2024-02' THEN revenue END) /
    NULLIF(SUM(CASE WHEN month = '2024-01' THEN revenue END), 0) - 1 AS change_pct
FROM orders
WHERE status = 'completed'
GROUP BY category, channel
ORDER BY change_pct ASC;

Step 4 — Recommend: Once you identify where the drop is concentrated (e.g., Electronics on mobile), investigate that segment deeply. Did pricing change? Did a competitor launch? Is there a technical issue (mobile checkout broken)?

---

Case 2 — Declining Conversion Rate

"Our website conversion rate dropped from 3.2% to 2.4%. What do you do?"

Show answer

Step 1 — Clarify: - What is the conversion definition? (Purchase? Sign-up? Add to cart?) - When did this change? (Gradual or sudden?) - Any recent changes — website redesign, new checkout flow, traffic source change?

Step 2 — Funnel analysis:

Traffic → Product View → Add to Cart → Checkout → Purchase
Before:  100%    62%         41%          31%       3.2%
After:   100%    58%         39%          29%       2.4%

The biggest relative drop is at Checkout → Purchase (31% → 29% = 6.5% relative decline).

Step 3 — Possible causes: - Payment method issue (credit card processor outage) - New required field at checkout causing abandonment - Price increase reducing conversion at final step - Traffic source changed (lower-intent users from a new channel)

Step 4 — Investigate: - Check payment error rates (technical) - Compare conversion by traffic source (marketing) - Review checkout session recordings (product) - A/B test checkout changes (if recent)

---

Case 3 — Customer Churn

"Our churn rate increased from 5% to 8% this quarter. How do you diagnose this?"

Show answer

Step 1 — Define churn: What counts as churned? (Cancelled subscription, no purchase in 90 days, explicitly cancelled?)

Step 2 — Segment who is churning:

-- Which customer segments are churning most?
SELECT segment, tenure_months_bucket, COUNT(*) AS churned
FROM customers
WHERE churned_this_quarter = 1
GROUP BY segment, tenure_months_bucket
ORDER BY churned DESC;

Is it new customers (poor onboarding?), long-tenured customers (competitor pulling them away?), a specific plan tier?

Step 3 — Diagnose the cause: - Product: Did engagement metrics (logins, feature usage) drop before churn? - Service: Did support ticket volume increase before churn? - Pricing: Did we raise prices? Did a competitor drop theirs? - External: Is this industry-wide (economic conditions)?

Step 4 — Intervention: - Early warning system: flag accounts whose usage has dropped 40% in the last 30 days - Proactive outreach by CSM to at-risk accounts - Exit survey: understand the stated reason for leaving

---

Case 4 — Metric Design

"Design a dashboard for the head of e-commerce operations. What would you put on it?"

Show answer

Step 1 — Ask about the audience: What decisions does the head of operations make daily? What would cause them to act?

Step 2 — Identify the 3-5 KPIs that matter most:

KPI	Why it matters	Alert threshold
Order fulfilment rate	% orders shipped on time	< 95%
Average delivery time	Customer satisfaction driver	> 3 days
Return rate	Indicates product quality issues	> 12%
Inventory stockout rate	% SKUs with 0 stock	> 5%
Cost per order	Operational efficiency	> £8

Step 3 — Dashboard layout: - Top row: 5 KPI cards with trend arrows (RAG status) - Middle: operational alerts (orders past SLA, stockouts by category) - Bottom: time series of fulfilment rate and delivery time

Step 4 — What to leave out: Revenue, margin, customer acquisition — those belong on a commercial dashboard, not operations.

---

Case 5 — A/B Test Design

"How would you test whether adding customer reviews to the product page increases conversion?"

Show answer

Step 1 — Hypothesis: H₀: Adding reviews has no effect on conversion rate. H₁: Adding reviews increases conversion rate.

Step 2 — Experiment design: - Control (A): product page without reviews - Treatment (B): product page with customer reviews section - Randomisation unit: user (not session — avoid contamination) - Split: 50/50 - Duration: minimum 2 weeks (to capture weekly seasonality)

Step 3 — Primary metric: Conversion rate (% of product page visitors who purchase)

Step 4 — Sample size calculation:

# If baseline conversion = 3.5%, minimum detectable effect = 0.5%
from statsmodels.stats.power import NormalIndPower
# ~4,800 users per group required (at 80% power, α=0.05)

Step 5 — Guardrail metrics: Average order value (do reviews change what people buy?), return rate (do reviews cause misleading expectations?), time on page (does reading reviews delay purchase?).

Step 6 — Analysis: Chi-square test on conversion proportions. Report absolute and relative lift with 95% CI.

---

← KPI & Metrics · Next: Dashboard Discussions →