Investigate ride declines and test free trials

Quick Overview

This question evaluates proficiency in product analytics, cohort and funnel analysis, retention and supply diagnostics, pricing and subscription economics, and experimentation design.

Investigate ride declines and test free trials

Company: Uber

Role: Data Scientist

Category: Analytics & Experimentation

Difficulty: hard

Interview Round: Onsite

Rides on Los Angeles shared vehicles are falling 10% month-over-month. Build a diagnostic and action plan: (a) Decompose the decline into acquisition (new riders), activation, cohort retention, and supply constraints using cohort tables and time-of-day/geo cuts; list the exact plots/queries you would run and what each would prove or falsify. (b) Enumerate plausible root causes (seasonality, price/ETA shifts, ops outages, competitor entry) and the minimal data to refute each. (c) For a '$100/month unlimited after a 1-week free trial' offer, decide who should receive the trial (new users, churn-risk users, high-propensity payers) and why; define targeting rules. (d) Provide the break-even formula for the trial: incremental LTV × conversion − CAC − variable costs − fraud loss ≥ 0; solve for the minimum conversion rate given inputs. (e) Outline the A/B test design (unit, duration, sample size, stratification), guardrails (support load, refund rate, fraud), and success criteria for a 4-week run.

Quick Answer: This question evaluates proficiency in product analytics, cohort and funnel analysis, retention and supply diagnostics, pricing and subscription economics, and experimentation design.

|Home/Analytics & Experimentation/Uber
Uber logo
Uber
Oct 13, 2025, 9:49 PM
hardData ScientistOnsiteAnalytics & Experimentation
10
0

LA Shared Rides Down 10% MoM — Diagnostic And Action Plan

Context: The Los Angeles market is seeing a 10% month-over-month decline in completed rides on shared vehicles. You are asked to diagnose the drop and design actions, including a promotional subscription test.

Tasks

(a) Decompose the decline into acquisition (new riders), activation, cohort retention, and supply constraints using cohort tables and time-of-day/geo cuts. List the exact plots/queries you would run and what each would prove or falsify.

(b) Enumerate plausible root causes (seasonality, price/ETA shifts, ops outages, competitor entry, etc.) and the minimal data to refute each.

(c) For a "$100/month unlimited after a 1-week free trial" offer, decide who should receive the trial (new users, churn-risk users, high-propensity payers) and why; define targeting rules.

(d) Provide the break-even formula for the trial: incremental LTV × conversion − CAC − variable costs − fraud loss ≥ 0; solve for the minimum conversion rate given inputs.

(e) Outline the A/B test design (unit, duration, sample size, stratification), guardrails (support load, refund rate, fraud), and success criteria for a 4-week run.

Loading comments...