Project Context
At DoorDash, you are the program manager supporting the launch of a new in-app courier incentive program across 6 U.S. cities. The launch decision is scheduled for next Friday, but two executive dashboards disagree on a key metric: one shows courier acceptance rate improved by 4.8%, while the other shows a 1.1% decline for the same pilot period. The team is 9 people across product, data, engineering, operations, and finance, and leadership needs a recommendation within 7 days.
Key Stakeholders
The GM of Logistics wants to launch on time to hit quarterly efficiency targets. Finance needs confidence in the projected $1.2M quarterly savings before approving rollout budget. The Data Science lead cares most about metric integrity and wants to delay if the discrepancy is unresolved. City Operations managers are already preparing courier communications and training materials.
Constraints
- Pilot ran for 21 days across 6 cities
- Rollout budget is capped at $350,000
- No extension of the pilot is approved
- 2 data engineers are each only available 30% due to a parallel warehouse migration
- Executive launch review is in 7 calendar days
- A marketing email to 42,000 couriers is scheduled to send 48 hours after the review unless canceled
Complications
- The discrepancy appears only in 2 of the 6 cities, and those cities account for 38% of pilot volume.
- One dashboard uses event-stream data, while the other uses the finance warehouse, and ownership is split across different teams.
- The VP of Operations has already informally told regional leaders the rollout is likely approved.
Deliverables
- Create a 7-day execution plan to investigate the discrepancy and support a go/no-go recommendation.
- Define how you would triage root causes, assign owners, and communicate daily status to stakeholders.
- Recommend launch options, including whether to proceed, delay, or phase rollout by city.
- Specify success criteria for resolving the discrepancy and making a defensible decision.
- Identify the top risks, mitigations, and rollback or contingency actions if the issue is not fully resolved by the review.
