PracHub
QuestionsPremiumCoachesLearningGuidesInterview Prep
|Home/Analytics & Experimentation/PayPal

Diagnose drop in shopper order acceptance

Last updated: Jun 15, 2026

Quick Overview

A PayPal data-scientist onsite analytics case: shopper order acceptance in a grocery-delivery marketplace drops by two-thirds on Sunday afternoon, and you must diagnose the cause. It tests precise metric definition and funnel decomposition, supply/demand/merchant root-cause hypotheses, metric-driven investigation, ruling out logging artifacts, and proposing mitigations plus a switchback experiment.

  • easy
  • PayPal
  • Analytics & Experimentation
  • Data Scientist

Diagnose drop in shopper order acceptance

Company: PayPal

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: easy

Interview Round: Onsite

##### Question Marketplace diagnosis case. A grocery-delivery marketplace (Instacart-style) observes that on **Sunday afternoon**, the number of orders that shoppers **accept** drops by about **2/3** compared to the usual baseline. Assume this is a same-day change (not a long multi-month trend). You have access to typical marketplace logs: order creation / checkout, dispatch and offer events, acceptances, cancellations, ETAs, shopper app events, and merchant/store signals. Acting as the on-call, bar-raiser-style interviewer, diagnose the problem. 1. **Clarify the metric.** Define precisely what "orders accepted" means and which denominator(s) matter (e.g. accepted count vs. acceptance rate = accepted / offered). State a clean funnel decomposition. 2. **Build a structured root-cause tree** from three perspectives: (a) **shopper supply / behavior**, (b) **customer demand / order mix**, and (c) **merchant / store operations and platform systems**. 3. **List the key metrics and slices you would check first** (at least 10), and state what pattern in each would support or refute a hypothesis. 4. **Distinguish a real behavioral change from a logging / measurement artifact** — how would you confirm the drop is real? 5. **Propose immediate mitigations and longer-term fixes**, including at least one experiment or controlled rollout plan to validate the leading hypothesis and prevent recurrence, and state the "next action" you would recommend.

Quick Answer: A PayPal data-scientist onsite analytics case: shopper order acceptance in a grocery-delivery marketplace drops by two-thirds on Sunday afternoon, and you must diagnose the cause. It tests precise metric definition and funnel decomposition, supply/demand/merchant root-cause hypotheses, metric-driven investigation, ruling out logging artifacts, and proposing mitigations plus a switchback experiment.

Related Interview Questions

  • How would you measure impact? - PayPal (medium)
  • How to evaluate a new homepage feature - PayPal (easy)
  • Design and evaluate a fraud detection strategy - PayPal (easy)
  • Design a fraud mitigation strategy under constraints - PayPal (hard)
  • Design metrics and experiment for donation feature - PayPal (easy)
PayPal logo
PayPal
Oct 10, 2025, 12:00 AM
Data Scientist
Onsite
Analytics & Experimentation
3
0
Question

Marketplace diagnosis case. A grocery-delivery marketplace (Instacart-style) observes that on Sunday afternoon, the number of orders that shoppers accept drops by about 2/3 compared to the usual baseline.

Assume this is a same-day change (not a long multi-month trend). You have access to typical marketplace logs: order creation / checkout, dispatch and offer events, acceptances, cancellations, ETAs, shopper app events, and merchant/store signals. Acting as the on-call, bar-raiser-style interviewer, diagnose the problem.

  1. Clarify the metric. Define precisely what "orders accepted" means and which denominator(s) matter (e.g. accepted count vs. acceptance rate = accepted / offered). State a clean funnel decomposition.
  2. Build a structured root-cause tree from three perspectives: (a) shopper supply / behavior , (b) customer demand / order mix , and (c) merchant / store operations and platform systems .
  3. List the key metrics and slices you would check first (at least 10), and state what pattern in each would support or refute a hypothesis.
  4. Distinguish a real behavioral change from a logging / measurement artifact — how would you confirm the drop is real?
  5. Propose immediate mitigations and longer-term fixes , including at least one experiment or controlled rollout plan to validate the leading hypothesis and prevent recurrence, and state the "next action" you would recommend.

Solution

Show

Submit Your Answer to Earn 20XP

Sign in to leave a comment

Loading comments...

Browse More Questions

More Analytics & Experimentation•More PayPal•More Data Scientist•PayPal Data Scientist•PayPal 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.