Context: High return rates were driving revenue erosion and customer dissatisfaction in Indonesia.

Problem: The existing workflow treated returns as a loss instead of a retention opportunity.

Decision: Coordinated the Indonesia rollout of an exchange feature that surfaced alternative inventory options before refund processing, aligning marketplace, support, and operations teams.

Result: Converted ~15% of returns into exchanges, preserving €695K in revenue.

Context: Shipping fees were misaligned with underlying logistics costs, creating margin leakage across transactions.

Problem: Flat discount logic failed to account for cost-per-kg structure, eroding unit economics at the SKU and lane level.

Decision: Diagnosed logistics subsidy leakage via SKU and lane-level analysis. Redesigned shipping fee models to realign pricing with cost structure within system constraints.

Result: Reduced annual costs by €520K without degrading service levels.

Context: Seller support coverage was limited to business hours, leading to SLA breaches and inefficient L1 handling.

Problem: Human availability constrained response times and created backlog during peak periods.

Decision: Integrated a third-party AI chatbot into seller support workflows via Intercom. Defined KPIs, structured the rollout plan, and monitored deflection performance across regions.

Result: Achieved 68% ticket deflection and expanded coverage to 24/7 autonomous support without additional headcount.

Context: Payment processing costs were fully absorbed by the platform while charging a handling fee had become standard practice across the e-commerce market, creating an unfavorable cost-income ratio.

Problem: Introducing a handling fee risked customer churn if rates were misaligned with market norms. The design required balancing cost recovery against customer experience — including whether to differentiate fees by payment method.

Decision: Modeled margin sensitivity and benchmarked handling fee structures across major regional platforms. Recommended a tiered structure with differentiated rates for standard and COD payments, staying within market range while recovering meaningful margin.

Result: Projected €200K annual revenue uplift. Proposal went live after handover to successor team.

Context: Route activation required manual configuration, creating a 5-day lead time and limiting network scalability.

Problem: Manual setup dependencies introduced delays and operational bottlenecks across a 130-node logistics network.

Decision: Designed the logic-architecture for a new nationwide routing engine, partnering with engineering to implement automated activation rules and remove manual configuration dependencies.

Result: Delivered a 60% reduction in activation lead time (5 days → 2 days), enabling faster network expansion.

Context: Regional seller onboarding workflows were fragmented across 6 markets, creating cross-functional friction and avoidable support ticket volume.

Problem: Inconsistent processes between regional and local teams limited scalability and created operational confusion.

Decision: Designed and deployed a unified regional operating model, consolidating fragmented workflows into a single global system with clarified ownership boundaries and codified operating procedures.

Result: Eliminated back-and-forth tickets between regional and local teams, enabling scalable onboarding across markets.