DoorDash — Dasher & Logistics — A/B Test & Causal Inference Questions

Role context: Data Scientist, Dasher & Logistics Analytics · Est. study time: 50 min · 4 questions

How experimentation works here

Logistics experimentation at DoorDash is defined by marketplace interference:

• Treatment and control orders in the same market share a Dasher fleet, so an order/user-level A/B is biased — the workhorse is the switchback (randomize a region × time window).
• Switchbacks trade bias for power and carry temporal autocorrelation and few effective units.
• Dasher supply responds to pay but can't be cleanly A/B'd, so supply effects come from causal designs.
• Logistics changes trade off on-time, cost, and Dasher earnings — the OEC is never a single metric.

For the fundamentals — p-values, power, error types, distributions — see the Probability & Statistics section.

Each answer is a coaching walkthrough: a Sample answer (clarify → approach → a simulated back-and-forth → a clear call), then a Deep dive with illustrative example, then a Grading rubric.