Vlad Severin

Search

SearchSearch
        • AB Testing infrastructure
        • Confidence interval
        • Experiment duration
        • False positives & False negatives
        • Hypothesis
        • Learning from experiment
        • What is AB testing
      • AB Test checklist
      • AB Test metrics
      • AB Test workflow
      • AB Testing process
      • AB Testing service (old)
      • AB Tests
      • AB Tests infrastructure
      • AB Tests prioritization
      • AB Tests solution
      • Feature toggles
      • Ideas for AB tests
      • Links
      • Problem statement
      • Stats Calculator
      • User segmentation
      • (item-based) Pearson similarity
      • Adjusted cosine similarity
      • Cosine similarity
        • Aggregate demand and aggregate supply
        • Central bank
        • Circular flow model
        • Economic growth
        • Equity in distribution of income
        • Fiscal policy
        • Gross Domestic Product (GDP)
        • Inflation
        • Interest rates
        • Level of economic activity
        • Low and stable inflation rate
        • Low unemployment
        • Macroeconomics
        • Monetary policy
        • Average reciprocal hit rate (ARHR)
        • Churn
        • Content-based recommendations
        • Coverage
        • Diversity
        • Evaluating recommender system
        • Hit rate
        • Mean absolute error (MAE)
        • Novelty
        • Recommender system
        • Responsiveness
        • Root mean square error (RMSE)
        • Top-N recommender
          • Average
          • Mean
          • Median
          • Pooled Standard Deviation
          • Standard deviation
          • True mean
          • Variance
          • Binominal distribution
          • Distribution
          • Normal Distrubution
          • Normality assumption
          • Sampling distribution
          • t-distrubution
          • Confidence intervals
          • Confidence level
          • Hypothesis testing basics
          • Minimal Effect of Interest (MEI)
          • one-sided vs two-sided
          • p-value
          • Rejection region approach
          • Significance level
          • Statistical power
          • Statistical significance
          • t-test
          • Test power
          • Type 1 and Type 2 errors
          • z-test
        • Dependent Variable (DV)
        • Simpson's Paradox
    Home

    ❯

    AB Tests

    ❯

    AB Test checklist

    AB Test checklist

    Feb 25, 2024, 1 min read

    Preparing for the test

    1. Choose primary and secondary metrics.
    2. Choose test power (1 - beta). Usually it’s 0.8
    3. Choose significance level (alpha). Usually it’s 0.05
    4. Obtain existing mean value (mean 1 or mu1) from historical data
    5. Chose MEI and calculate mean 2 (mu2)
    6. Calculate sample size using stats calculator
    7. Based on sample size calculate test duration

    Graph View

    Backlinks

    • AB Test workflow

    Created with Quartz v4.2.2 © 2024

    • GitHub