Repro - Mobile Analytics for growth
日本語
Sign Up Back to Dashboard
  • System Requirements
  • Dashboard Guide
  • Development Guide
  • Release Notes
  • FAQ
    • App
    • Web
    • General
      • Analytics
      • A/B test significant difference display
        • At what point is the data collected that is used for computing the significant difference?
        • The screen hasn’t changed from “There isn’t a significant difference yet”. Is there a rough indication of when it will be completed?
        • Around how many delivery counts is needed to display the results?
        • What is a significant difference?
        • What is a confidence interval?
        • Will the results change if it was determined to have a significant difference?
        • Is it possible to go back to past significant differences?
        • Is it possible to change the significance level used in the confidence interval?
        • When is automatic winner selection performed?
        • Can I enable automatic winner selection for a campaign that is already being delivered?
        • Can the control group be automatically selected as the winning variant?
      • API
      • Miscellaneous

What is a significant difference?¶

It is a statistical judgment of whether the difference in the target indicator (such as CVR, open rate, or button tap rate) is not due to chance. The calculation uses Always Valid Confidence Sequences, a sequential testing method designed so that the probability of a false judgment does not increase even with repeated evaluations. For details, see for example Statistical inference.

  • « Around how many delivery counts is needed to display the results?
  • What is a confidence interval? »

Repro Service Site About Us Careers Terms of Service Privacy Policy Cookie Policy

© 2022 Repro Inc.