Back to all cases
CASE STUDY / Data Analytics
Cancellation Timing & Order Lifecycle Analysis
O2O On-Demand Logistics PlatformsAnalyzed when cancellations occurred across order states, separating pre- and post-match behavior to identify the highest-value intervention windows.
Delivery Details
This analysis extended cancellation rate from a single outcome metric into a time-to-event study, calculating intervals from booking to cancellation, matching to cancellation, and acceptance to cancellation while building state-based timing distributions and cumulative curves. Segmentation by time bucket, cancelling party, wait duration, match state, time of day, market, and cohort distinguished immediate cancellations, long-wait cancellations, and post-match cancellations rather than treating them as one behavior. Findings informed early-warning indicators, operational intervention windows, and more precise cancellation monitoring dashboards.
Industry Context
Cancellation risk in on-demand services changes quickly with order state and waiting time. Pre-match cancellation, cancellation after a long wait, and post-acceptance cancellation represent different experience and operational failures, requiring a state- and time-aware view rather than one final rate.
Business Problem
The same cancellation outcome could occur at very different order stages, so the rate alone could not show when intervention was needed.
What I Built
Built order-state and time-to-event analysis across timing windows, cancelling parties, and waiting conditions.
Business Impact
Identified stages where cancellation risk began to rise, allowing product and operational actions to target the right intervention moments.
For career conversations, advisory work, or a concrete Data & AI project, feel free to get in touch.