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

Experiment Speed vs Rigor Tradeoff

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

StreamCart, a grocery delivery app, wants to test a faster checkout flow. The product manager argues for shipping after one week to increase experimentation velocity, while the data science team wants enough sample size to maintain statistical power.

Problem Statement

You need to quantify the trade-off between running a short experiment and waiting for a statistically rigorous result. Assume the team cares about detecting a meaningful lift in checkout conversion.

Given Data

MetricValue
Baseline checkout conversion rate18.0%
Minimum detectable effect (absolute)1.5 percentage points
Significance level0.05
Desired power80%
Daily eligible users40,000
Traffic split50% control / 50% treatment
One-week observed control users140,000
One-week observed treatment users140,000
One-week observed control conversions25,200
One-week observed treatment conversions27,020

The one-week result implies treatment conversion of 19.3% versus 18.0% in control.

Requirements

  1. Compute the required sample size per group to detect the target 1.5 percentage point lift with 80% power.
  2. Convert that sample size into the approximate number of days needed at the given traffic level.
  3. Using the one-week observed data, run a two-proportion z-test.
  4. Compute a 95% confidence interval for the observed lift.
  5. Decide whether shipping after one week is justified.
  6. Explain how you would balance experimentation speed and statistical rigor in practice.

Assumptions

  • Users are independently randomized at the user level.
  • No major novelty effect or traffic-quality shift occurs during the test.
  • The normal approximation for proportions is valid.
  • Use a two-sided test.