PracHub
QuestionsPremiumCoachesLearningGuidesInterview Prep
|Home/Analytics & Experimentation/DoorDash

Investigate LA Order Drop

Last updated: Apr 11, 2026

Quick Overview

This question evaluates marketplace analytics, causal inference, hypothesis-driven investigation, funnel decomposition, and experimentation design skills in a data science context.

  • medium
  • DoorDash
  • Analytics & Experimentation
  • Data Scientist

Investigate LA Order Drop

Company: DoorDash

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: medium

Interview Round: Technical Screen

A food delivery marketplace sees a meaningful decline in completed orders in the Los Angeles market. Explain how you would investigate the drop end to end. Your answer should: 1. Break the problem into major components such as: - customer demand - merchant supply - courier supply - product funnel and reliability - external factors such as seasonality, weather, pricing, promotions, and competitor activity 2. List concrete hypotheses for each component and describe what data you would use to confirm or reject each hypothesis. 3. Decompose the order decline across the funnel, for example: - app visits or sessions - search - menu views - add-to-cart - checkout - order attempts - completed orders Also account for marketplace constraints such as merchant online minutes, delivery radius, ETA, stockouts, acceptance rate, cancellations, and operational incidents. 4. Suppose your investigation suggests that the main driver is that merchants are online for fewer minutes per day. The business now wants to test an incentive program intended to increase merchant online time. Design an experiment. For the experiment design, address: - the appropriate randomization unit - why a merchant-level switchback design is not appropriate if a merchant cannot realistically be treated and untreated within the same day - whether a geo-level or cluster-level holdout is better, and how you would reduce interference - primary, secondary, and guardrail metrics - eligibility rules, treatment assignment, and experiment duration - how to perform power analysis or minimum detectable effect planning - how you would interpret the result if merchant online minutes increase but completed orders do not

Quick Answer: This question evaluates marketplace analytics, causal inference, hypothesis-driven investigation, funnel decomposition, and experimentation design skills in a data science context.

Related Interview Questions

  • Evaluate Biker Feature Success - DoorDash (hard)
  • How would you test product changes? - DoorDash (hard)
  • How to test bike delivery? - DoorDash (medium)
  • Investigate LA successful orders drop - DoorDash (easy)
  • How would you diagnose a completed orders drop? - DoorDash (easy)
DoorDash logo
DoorDash
Jan 18, 2026, 12:00 AM
Data Scientist
Technical Screen
Analytics & Experimentation
6
0

A food delivery marketplace sees a meaningful decline in completed orders in the Los Angeles market. Explain how you would investigate the drop end to end.

Your answer should:

  1. Break the problem into major components such as:
    • customer demand
    • merchant supply
    • courier supply
    • product funnel and reliability
    • external factors such as seasonality, weather, pricing, promotions, and competitor activity
  2. List concrete hypotheses for each component and describe what data you would use to confirm or reject each hypothesis.
  3. Decompose the order decline across the funnel, for example:
    • app visits or sessions
    • search
    • menu views
    • add-to-cart
    • checkout
    • order attempts
    • completed orders
    Also account for marketplace constraints such as merchant online minutes, delivery radius, ETA, stockouts, acceptance rate, cancellations, and operational incidents.
  4. Suppose your investigation suggests that the main driver is that merchants are online for fewer minutes per day. The business now wants to test an incentive program intended to increase merchant online time. Design an experiment.

For the experiment design, address:

  • the appropriate randomization unit
  • why a merchant-level switchback design is not appropriate if a merchant cannot realistically be treated and untreated within the same day
  • whether a geo-level or cluster-level holdout is better, and how you would reduce interference
  • primary, secondary, and guardrail metrics
  • eligibility rules, treatment assignment, and experiment duration
  • how to perform power analysis or minimum detectable effect planning
  • how you would interpret the result if merchant online minutes increase but completed orders do not

Solution

Show

Submit Your Answer to Earn 20XP

Sign in to leave a comment

Loading comments...

Browse More Questions

More Analytics & Experimentation•More DoorDash•More Data Scientist•DoorDash Data Scientist•DoorDash Analytics & Experimentation•Data Scientist Analytics & Experimentation
PracHub

Master your tech interviews with 8,000+ real questions from top companies.

Product

  • Questions
  • Learning Tracks
  • Interview Guides
  • Resources
  • Premium
  • For Universities
  • Student Access

Browse

  • By Company
  • By Role
  • By Category
  • Topic Hubs
  • SQL Questions
  • Compare Platforms
  • Discord Community

Support

  • support@prachub.com
  • (916) 541-4762

Legal

  • Privacy Policy
  • Terms of Service
  • About Us

© 2026 PracHub. All rights reserved.