As asked
DoorDash has expanded into grocery delivery through DashMart and partnerships with retailers like Walmart, Kroger, and Petco. From a backend engineering perspective, what are the top three risks you would want to prototype or de-risk before a grocery launch in a new metro, and how would you approach each?
Sample answer outline
Strong answers identify domain-specific risks over generic engineering risks. Good candidates name: real-time inventory availability (grocery items go out of stock mid-shopping, requiring substitution logic that restaurant orders do not need, and DoorDash surfaces this through a Dasher in-app substitution flow), item-level weight pricing (bananas are priced by weight, not unit count, so the order total cannot be finalized until pickup and the authorization hold must cover the upper bound), and the fulfillment model difference (a Dasher shops the store for 10 to 30 minutes versus picking up a ready order, fundamentally changing dispatch timing and reassignment logic). Each risk needs a specific prototype or integration test plan.
Expect these follow-ups
- How would you design the substitution flow where a Dasher can suggest a replacement for an out-of-stock item?
- What metrics would you track in the first 30 days to know if the launch is healthy?