Dataford
Popular roles
Software EngineerData AnalystData ScientistData EngineerBusiness AnalystAI EngineerMachine Learning EngineerProduct Manager
Browse
Browse All RolesEvery role hub, from analyst to MLBrowse All CompaniesCompany-specific interview loopsAll Interview GuidesThe full guide library
Top questions by role
Software EngineerData AnalystData ScientistData EngineerBusiness AnalystAI EngineerMachine Learning EngineerProduct Manager
Top questions by skill
SQLPythonStatisticsMachine LearningA/B TestingSystem DesignGenerative AIProduct SenseMetricsBehavioral
Browse all questions →Try a mock interview
ExperiencesCoaching
New
Practice
Mock InterviewsTimed interview simulations with feedbackSuccess PathYour 6-week structured planModulesCurated lessons by topicWebinarsTalks from ex-Big Tech data leadsPlaygroundA free-form scratch editor
Learn
BlogInterview strategy and career adviceTech Job Market ReportHiring trends across data and AI rolesFor UniversitiesDataford for career centersAbout DatafordWho we are and how we build
Pricing
Get started

Correlation vs Causation in Growth Campaigns

MediumStatistics & Probability00:00
Practice interviewer
In session
5 left
00:00

Whenever you're ready, walk me through how you'd approach this.

You need to log in / sign up to chat or submit.

Problem

Business Context

FitPulse, a subscription fitness app, noticed that users who received more push notifications tended to have higher 30-day retention. The growth team wants to know whether notifications caused the retention lift or whether the relationship is only correlational.

Problem Statement

You are given observational data split by user engagement segment. At the aggregate level, users who received notifications appear to retain better. Determine whether that alone supports a causal claim, and quantify the difference between the naive aggregate conclusion and the segment-adjusted conclusion.

Given Data

SegmentNotification Group UsersRetained in Notification GroupNo-Notification UsersRetained in No-Notification Group
High engagement8,0005,6002,0001,500
Low engagement2,0004008,0001,200

Additional information:

MetricValue
Significance level0.05
Total notification users10,000
Total no-notification users10,000

Requirements

  1. Compute the aggregate 30-day retention rate for notification vs no-notification users.
  2. Test whether the aggregate difference is statistically significant using a two-proportion z-test.
  3. Compute retention rates within each engagement segment.
  4. Explain why the aggregate relationship does not imply causation.
  5. Estimate the notification effect after adjusting for engagement segment using a weighted average of within-segment differences.
  6. State the business conclusion: should FitPulse claim notifications caused retention to improve?

Assumptions

  • Retention is measured consistently across all users.
  • Engagement segment is a pre-existing confounder known before notifications were sent.
  • This is observational data, not a randomized experiment.
  • Ignore other confounders beyond engagement for this exercise.