Katkat

Investigate LA successful orders drop

DoorDashProduct Analyst · Technical Screen

Last updated

This question tests your ability to methodically decompose a performance metric drop, isolate potential root causes using data dimensions, and suggest analytical steps to distinguish between technical issues and marketplace dynamics.

Scenario

You are a Product Analyst at DoorDash. Your dashboard shows that the total number of 'Successful Orders' in the Los Angeles market dropped by 15% yesterday compared to the same day in the previous week. Your manager has asked you to investigate the root cause.

You have access to the following tables:

  • orders: order_id, timestamp, status (completed, cancelled, failed), customer_id, delivery_zone.
  • events: order_id, event_type (app_open, restaurant_search, checkout_attempt, payment_success), timestamp.
  • operations: order_id, dasher_id, assignment_time, pickup_time.

Task

  1. How would you structure your investigation to identify the root cause? Provide a step-by-step framework.
  2. What specific segments or cohorts would you check to determine if the issue is localized to a specific group of users, restaurants, or technical platforms?
  3. If you find the issue is related to a high rate of 'checkout_attempt' to 'payment_success' failure, what metrics would you calculate to narrow down the cause?

Example

Example of a diagnostic hypothesis:

  • Observation: A drop in successful orders is observed.
  • Split by platform: If the drop is only on iOS, it suggests a recent app update issue.
  • Split by zone: If the drop is only in a specific zip code, it might indicate a local infrastructure issue (e.g., severe weather or a service outage).

Hints

Hint 1

Start by ensuring the drop is an actual trend rather than a reporting lag or data pipeline error.

Hint 2

Perform a funnel analysis to see which specific stage of the order flow (from search to delivery) is experiencing the highest conversion leakage.

Hint 3

Consider external factors that could impact the Los Angeles market, such as local events, changes in delivery fees, or competitor activity.

Follow-up questions

  • If the data shows that successful orders dropped because of an increase in order cancellations by customers, what product features would you investigate?
  • How would you verify if the drop in orders is due to an insufficient supply of Dashers in the area?
  • If you identify a bug in the payment gateway as the cause, how would you design an experiment to measure the impact of the fix once implemented?