DoorDash — Dasher & Logistics — Product Sense & Metrics
Role context: Data Scientist, Dasher & Logistics Analytics · Est. study time: 30 min · 3 practice questions
How to prepare for this role
DoorDash Dasher & Logistics Analytics is the most experimentation-and-causal-inference-heavy team: dispatch, ETA, batching, Dasher supply, and delivery quality. Its signature is that logistics changes can't be user-randomized — treatment and control orders share the same Dasher fleet — so DoorDash relies on switchback experiments.
The day-to-day is measuring and optimizing the delivery/supply side: matching orders to Dashers, delivery times, batching, and supply-demand balance per market. The instinct that separates strong candidates: knowing an order-level A/B is biased in logistics (shared supply) and reaching for a switchback, and framing the objective as reliable on-time delivery at sustainable cost — not minimum delivery time.
Where to spend your prep time:
- Logistics product sense (this article) — the delivery-time/cost/Dasher-pay triangle, batching, and supply-demand balance.
- Switchback experimentation and causal inference — why user-level A/B is biased, switchback design and its pitfalls, Dasher-supply elasticity. See the A/B Test & Causal Inference section.
- SQL and communication — be fluent in SQL, and turn a supply or dispatch analysis into a clear recommendation.
The through-line: balance on-time delivery, cost, and Dasher supply, and measure logistics changes with switchbacks because they interfere through shared supply.
What DoorDash Dasher & Logistics actually is
DoorDash fulfills ~776M orders a quarter with ~2M+ active Dashers. The identity to carry into every answer: a real-time logistics marketplace whose job is to match every order to a Dasher and deliver it fast, reliably, and at a sustainable cost, while keeping Dashers paid enough to keep showing up. The Dispatch system powers fulfillment across consumers, Dashers, and merchants.
Two things make this different from a normal analytics job:
- Everything trades off, three ways. Faster delivery needs more Dashers (cost) or less batching; more batching (multiple orders per Dasher) cuts cost but slows delivery; higher Dasher pay lifts supply but raises cost. The objective is reliable, on-time delivery at a sustainable cost with enough supply — not minimizing any one metric.
- Logistics changes interfere through shared supply. An order-level A/B is biased because treating some orders changes Dasher availability for the control orders in the same market (a network effect). So the unit of experimentation is often a region × time window — a switchback — not an order or user.