Uber Dispatch and ETA: A Systems Teardown

Tech · 7 min read

Uber's dispatch is a tightly coupled problem of routing, supply estimation, and user expectations. The system uses a mix of real-time telemetry and predictive models to match nearby drivers to riders while minimizing deadhead time. Constraints include driver preferences, surge pricing signals, regulatory rules, and the need to maintain low match latency to keep acceptance rates high.

ETA predictions are computed using historical travel times, live traffic, driver behavior models, and route uncertainty. Presenting a single ETA value involves smoothing and calibrating to avoid oscillation; unrealistic precision erodes trust. Gradual updates, confidence bands, and explicit caveats for traffic-sensitive trips are techniques Uber uses to manage expectations.

The UX around surge and ETA balancing has product and policy implications. Surge multipliers must be communicated clearly without driving away riders, and driver dispatch priority can impact perceived fairness. For product teams, the teardown underscores that matching algorithms must be paired with clear, empathetic communication to manage asymmetric patient tolerances between riders and drivers.