Causal Analysis Reveals Root Causes of Food Delivery Delays

A comprehensive causal analysis of over 405,000 food delivery orders in China reveals that pickup duration and transport time are the strongest drivers of delivery delays, rather than merely correlated factors. The research introduces a novel framework combining Bayesian causal discovery with double machine learning, distinguishing true causal relationships from statistical correlations. Critically, the study documents delay propagation: when preceding orders experience delays, subsequent orders handled by the same courier face compounded delays, creating a cascading effect throughout delivery waves.

With the OFD market projected to grow USD 470.5 billion through 2029 (26.9% CAGR), reducing delays directly impacts platform profitability, customer retention, and courier safety. Current delay rates of 38-44% for major platforms represent a systemic operational challenge that demands data-driven intervention. Supply chain leaders should prioritize targeting root causes rather than symptoms: optimizing merchant-to-courier handoff processes, implementing delay prediction systems that account for order wave dynamics, and redesigning courier assignment algorithms to prevent bottlenecks.

This research establishes a methodological blueprint for moving beyond predictive analytics to actionable causal inference. Organizations adopting similar frameworks can identify leverage points for operational improvement, translate causal insights into targeted interventions, and create more resilient delivery networks that enhance both economic and social sustainability.