Customer and User Obsession Questions

Compute: CAC = total acquisition and onboarding costs for enterprise feature / number of new enterprise customers acquired for that feature. LTV = sum of expected gross margin per customer over lifetime; for subscription: LTV = ARPA * gross margin % / churn rate. Payback period = CAC / monthly gross margin contribution. Unique enterprise considerations: longer sales cycles, higher CAC per account, upsell/expansion revenue, contract churn cliffs, and lower volume with higher variance. Use cohort-based LTV, include implementation/CS costs, and model expansion revenue (account growth). Sensitivity analysis: vary churn and expansion rates. Decision rules differ: longer acceptable payback and higher CAC tolerated if expansion and low churn yield strong LTV; account concentration risk and contract terms (multi-year, SLAs) must be modeled separately.

Framework: deliver a 48-hour diagnostic with clear steps, required data, and a concise deliverable. Steps: 1) Clarify cohort windows (signup week/month) and revenue definition (gross, net, ARR). 2) Pull data: user_id, cohort_date, revenue_events (amount, date), subscriptions, refunds. 3) Build cohort table: sum incremental revenue per user in first 90 days, then aggregate by cohort and segment (channel, plan). 4) Compute per-cohort metrics: total incremental revenue, revenue per user, conversion rate, and sample size/confidence intervals. Deliverable: a 1-page report + CSV with cohort breakdown and a short recommendation: top 3 cohorts by revenue-per-user and suggested experiments. I’d include SQL snippets, basic charts, and caveats about attribution and small sample sizes.

Situation: A product launch raised conversion by 4% but customer-service (CS) tickets doubled. Task: Quantify whether net business value justifies higher support costs and propose next steps. Action: 1) Compute incremental revenue: baseline revenue R_base, users U, conversion uplift ΔC=4% ⇒ ΔRevenue = U * AOV * ΔC. 2) Compute incremental support cost: ΔTickets = Tickets_post - Tickets_pre; convert to cost using Cost_per_ticket and marginal SLA staffing cost. 3) Calculate net benefit = ΔRevenue - ΔSupportCost and ROI = net / ΔSupportCost. 4) Segment analysis: compute uplift and ticket increase by cohort (new vs returning, device, geography) to find concentrated pain points. 5) Root-cause analysis on tickets (NLP topic modeling, entity extraction) to prioritize fixes. 6) Experiments: roll back or A/B test mitigations (UX tweaks, clearer copy, proactive help) on target cohorts and measure conversion + tickets. Result: If net benefit positive and concentrated in fixable cohorts, prioritize targeted fixes; if negative, rollback or redesign. Learnings: trade-offs must include marginal support cost and customer lifetime value, not just immediate conversion.

Checklist before deploying differential pricing personalization: 1) Fairness audit — test for protected attribute proxies and disparate impact across demographics. 2) Transparency & consent — ensure privacy policy and user communication about personalized offers. 3) Legal/compliance review — verify against regional pricing discrimination and consumer protection laws. 4) Data quality and bias — validate training data representativeness and absence of label bias. 5) Economic safeguards — cap price differentials and implement minimum/maximum bounds to avoid exploitation. 6) Customer experience & reputational risk — run holdout experiments and monitor CS feedback and social signals. 7) Monitoring & rollback — real-time metrics for complaints, conversion anomalies, and automated kill-switch if adverse outcomes appear.

Customer and user obsession means prioritizing real user needs and experiences above internal metrics or technical elegance — using data to deeply understand who users are, what problems they face, and delivering measurable value that improves their outcomes and satisfaction.

Three concrete behaviors and day-to-day examples:

1. Start with user questions, not metrics

• Behavior: Frame analyses around specific user problems or hypotheses; ask “what decision will this insight drive for the user?”
• Example: Before building retention models I meet the PM to clarify which user segment’s churn they care about and what action (e.g., targeted onboarding) will follow, then tailor features and labels accordingly.

2. Validate findings with qualitative feedback

• Behavior: Combine quantitative signals with user research (surveys, session recordings, interviews) to avoid misleading correlations.
• Example: After detecting a drop in feature usage, I pull session replay clips and run a short in-app survey to confirm whether UI confusion or a backend bug caused the decline.

3. Deliver actionable, measurable recommendations

• Behavior: Translate analyses into clear, testable product changes and define success metrics; partner to run experiments and iterate.
• Example: I propose a variant for an onboarding flow, implement an A/B test, and track conversion lift and downstream retention—then present results with suggested next steps.

These behaviors ensure data work consistently improves real user outcomes.