Apple Data Analyst Case Interview — App Store Subscriptions Growth and Experimentation

This Apple-style case simulates partnering with the App Store Services team to diagnose a slowdown in subscription growth and design a privacy-conscious experiment. You will be given a short product brief, a minimal schema, and noisy, aggregated datasets (no PII) reflecting Apple’s privacy-first approach. The interview emphasizes Apple’s culture of deep technical rigor, precise communication, and end‑user focus. What the case covers: 1) Problem framing and metrics definition: Clarify the business goal (increase net subscriber adds without harming customer experience). Define and justify KPIs/guardrails (e.g., trial start rate, paid conversion, churn, retention by cohort, refund rate, LTV, latency, App Store ratings impact), calling out seasonality around major product launches. 2) Data interrogation and quality: Assess data health across tables such as subscription_events (anonymized starts, conversions, cancellations, refunds), app_store_transactions (aggregated line items), device_telemetry_agg (weekly active devices by OS/hardware family), and experiment_log (assignment, variant, timestamp). Identify anomalies (e.g., attribution window changes, app version rollout, regional holidays), duplicate events, and how differential privacy noise or sampling impacts small cohorts. 3) Root-cause analysis: Form hypotheses for the growth slowdown (pricing page changes, paywall copy, search placement, payment declines, OS upgrade funnel friction, regional catalog shifts). Propose a quick metrics deep dive (cohort retention curves, pre/post change analysis, decomposition of net adds into acquisition vs. churn). 4) Experiment design under constraints: Design an A/B test on a new App Store subscription offer card. Define the primary metric (paid conversion within 14 days) and guardrails (refund rate, support contacts, crash rate). Calculate required sample size at Apple-scale, specify stratification (region, device family, OS), and discuss holdback vs. ramp strategy. Address privacy and platform constraints (aggregated logging, limited attribution). 5) Communication and storytelling: Present a crisp, executive-ready narrative: problem, insights, trade-offs, and decision. Provide a one-slide readout structure Apple teams expect (goal, key metrics deltas with CIs, risk/guardrails, recommendation, and next steps). 6) Execution details: Outline a minimal dashboard (north-star metric, drilldowns by country/device/OS/version, experiment panel, anomaly alerts). Draft pseudo-SQL to compute trial→paid conversion and churn by cohort; explain how you’d validate results (falsification checks, parallel trend checks, power re-estimation). Evaluation focuses on: structured thinking, attention to detail, data craftsmanship, respect for privacy-by-design, and the ability to balance product intuition with statistical rigor—hallmarks of Apple’s interview style.