Project Background
You are the program lead for ShopSphere, a global e-commerce marketplace (similar to Amazon Marketplace) with 45M monthly active users, 3.2M daily orders, and operations in 6 priority markets (US, Canada, UK, Germany, Japan, Australia). ShopSphere’s leadership believes there is meaningful upside in improving conversion and margin through pricing experimentation and personalization—but the company has historically used only rule-based pricing (category-level discounts, seasonal promos) and has never shipped user-level price personalization.
The CEO has committed to investors that ShopSphere will deliver a measurable improvement in profitability this half. The VP of Monetization is pushing a new initiative: “Smart Offers”—a system that can run controlled experiments on pricing and promotions and, in later phases, personalize offers based on user behavior (e.g., loyalty status, price sensitivity, cart abandonment signals). The immediate goal is not to deploy a fully automated ML pricing engine, but to launch a safe, compliant experimentation platform + first set of pricing/promo experiments that prove incremental value.
You have 10 weeks to deliver a launch that can run at least two experiments end-to-end and scale to more. The cross-functional team is partially staffed: 6 backend engineers (pricing + checkout), 2 data scientists (experimentation + causal inference), 1 applied scientist (personalization models), 2 analysts, 1 designer, 1 QA lead, and shared support from Legal/Privacy, Customer Support, and Finance. The platform must integrate with existing systems: a legacy pricing service (highly coupled), a promotions engine, and the checkout service. Any pricing mistake is high risk—ShopSphere processes $9B GMV/quarter, and pricing incidents have previously triggered social media backlash and regulatory complaints.
Stakeholder Landscape
- VP of Monetization (Exec Sponsor): Wants a visible win this quarter—ideally a conversion lift and margin improvement that can be attributed to Smart Offers. Prefers bigger, bolder tests (e.g., 5–10% price deltas) to show impact.
- Head of Trust & Compliance: Concerned about perceived “price discrimination,” fairness, and regulatory exposure (EU consumer protection, UK CMA scrutiny, and emerging US state privacy laws). Wants strict guardrails, audit logs, and clear customer messaging.
- Director of Engineering (Pricing/Checkout): Owns uptime and incident rate. Wants minimal changes to the legacy pricing service and insists on a robust rollback plan. Competing priority: a parallel project to migrate checkout to a new tax calculation provider.
- Finance Lead: Needs clean attribution and revenue recognition logic. Worried that experiments will distort forecasting and promo liability accounting.
- Country Managers (DE/JP): Want local control and worry that global experiments will break market-specific pricing norms (VAT-inclusive display in DE, consumer expectations in JP). They will block launches that risk reputational damage.
Constraints
| Constraint | Details |
|---|
| Timeline | 10 weeks to first production experiments; exec review in week 11 |
| Markets | Must launch in US + UK first; optional expansion to DE/JP if safe |
| Traffic allocation | Max 10% of sessions in treatment initially; must ramp gradually |
| Guardrails | No more than 0.5% increase in refund rate; no more than 0.3% increase in customer support contacts per order |
| Pricing limits | Per-item price changes capped at ±3% without VP approval; must respect MAP (minimum advertised price) constraints for 12 key brands |
| Tech constraints | Legacy pricing service supports only category-level rules; user-level overrides require a new “offer overlay” layer in checkout |
| Data constraints | Event logging is inconsistent across web and mobile; mobile app releases require 2-week store review windows |
| Budget | $150K for external legal review + additional monitoring tools; no additional headcount approved |
Deliverables (What you must produce in the interview)
- A prioritized roadmap for the 10-week window that sequences platform work, experiment design, and market rollout.
- A trade-off proposal: what you will cut or defer to hit the deadline (e.g., full personalization vs. segmented offers, mobile vs. web-first, number of markets).
- A launch plan including ramp strategy, monitoring, rollback criteria, and incident ownership.
- A stakeholder alignment plan: how you will get Legal/Trust, Finance, and Country Managers to sign off with clear decision points.
- A risk register with mitigations for the top risks (technical, regulatory, reputational, and measurement).
Complications (Assume these happen mid-project)
- Week 3: Legal flags that “personalized pricing” language is high risk in the EU/UK and requests you reframe the first launch as “personalized promotions” unless you can prove non-discrimination and provide customer-facing disclosures.
- Week 5: The checkout migration project pulls 2 backend engineers for an urgent tax provider cutover, reducing your engineering capacity by ~30% for 3 weeks.
- Week 7: A major brand partner (top-10 GMV) threatens to terminate the contract if any experiment violates MAP pricing—even accidentally—after they see a competitor get bad press for dynamic pricing.
Your Task
Walk through how you would execute this program end-to-end. Be explicit about:
- How you define MVP scope for experimentation vs. personalization.
- How you structure decision gates (e.g., legal approval, instrumentation readiness, pre-launch QA, go/no-go).
- How you ensure measurement integrity (consistent assignment, logging, and attribution) without boiling the ocean.
- How you handle the resource hit in week 5 and still deliver meaningful outcomes by week 10.
- What you will report to executives weekly (status, risks, and leading indicators).
