Optibase
  • ๐Ÿค”What is Optibase
    • ๐ŸงชWhat is a Test/Variant Group
    • ๐ŸชžWhat is a Variant
    • ๐ŸWhat is a Conversion
    • ๐Ÿง‘What is a Tested User
    • ๐Ÿ What is a Guest User
    • โš™๏ธWhat does the custom script do?
    • ๐Ÿ‘‘Probability to Be Best (P2BB)
  • ๐Ÿ‘ทHow to setup Optibase
    • โœ๏ธAdd the custom script
    • ๐ŸŽฏCreate a Conversion
    • ๐Ÿ”—Connect a Conversion
    • ๐Ÿ–ฅ๏ธServer-side Conversions
    • โš–๏ธCreate a Test/Variant Group
    • ๐Ÿ†ŽCreate an A/B Variant
    • ๐Ÿ”—Connect an A/B Variant
    • ๐Ÿ”€Create a Split Variant
    • ๐Ÿ”’Force Variants
    • ๐Ÿ‘‘Turn off when a winner is found
    • ๐Ÿ”žSet up restrictions
    • โฒ๏ธSchedule a Test
    • ๐Ÿ“จShare test results
    • ๐Ÿ”„Connect Conversions to a Test
    • ๐Ÿ“…Filter test results by a specific time period
  • ๐Ÿช„Optibase Webflow App
    • โœ๏ธAdd the custom script
    • ๐ŸŽฏCreate a Conversion
    • ๐Ÿ”—Connect a Conversion
    • โš–๏ธCreate a Test/Variant Group
    • ๐Ÿ†ŽCreate an A/B Variant
    • ๐Ÿ”—Connect an A/B Variant
    • ๐Ÿ”€Create a Split Variant
    • โ˜‘๏ธTypes of Content You Can Connect
  • ๐Ÿ”—Integrations
    • ๐Ÿ“žActive Variants API
    • โšกTrigger Conversions with GTM
    • ๐Ÿ“ˆGoogle Analytics 4
    • ๐Ÿ‘ทGoogle Analytics 4 with GTM
    • โœ–๏ธMixpanel
    • ๐Ÿฆ”Track conversions in PostHog
  • โ“FAQs
    • Why can you see views in the views tab for pages that don't have a test?
    • Why can't I see detailed metrics like average "Time on Page" and "Scroll Depth"?
    • Why do conversions in the Webflow app appear to disconnect randomly, and are these conversions still
    • Troubleshooting Component Connection Issues in Optibase
    • How Optibase Works with CMS in Webflow
    • Why Your Screen Flickers Before Loading the Right Variants
    • Why variants appear disconnected when inside a component?
    • Localisation Limitations in Webflow with Optibase
    • Why Conversions May Not Be Recording in Optibase
    • Does a URL Split Test Work with UTM Parameters?
    • Using Optibase Across Multiple Websites
    • Why Conversions or Results May Not Be Visible in the Results Tab
    • Can I redirect traffic between two different URLs using Optibase?
    • How do I exclude internal traffic from being counted in Optibase tests?
    • Can I connect multiple conversions to a single test?
    • How does Optibase handle geo-restricted A/B tests?
    • Why are conversion events double counted in my test?
    • How can I transfer ownership of an Optibase workspace?
    • Why do I see flickering between A/B variants?
    • Can I test elements across all pages on my website?
    • Why is a guest user still on a free plan after being invited to a paid workspace?
    • Can I redirect traffic between two different URLs using Optibase?
    • How do I set up a multivariate test in Optibase?
    • How do I exclude internal traffic from being counted in Optibase tests?
    • Why are conversion events double counted in my test?
    • Why is the wrong variant showing when I preview using the variant URL?
    • How can I force a specific variant to display for debugging?
    • How do I check which buttons are connected to a conversion event?
    • Can I run multiple tests on the same page?
    • Is it possible to set traffic allocation dynamically during a test?
    • How can I test different CTAs that exist across 50+ pages?
    • What happens when a test reaches statistical significance?
    • How do I view which conversion event was triggered?
    • Can I filter out specific IPs or users from test results?
    • Can I run tests on a site that uses a custom CMS or codebase?
    • Is Optibase compliant with GDPR?
    • Do I need to connect a conversion event to a test for it to be tracked?
    • How do I test different product pages with different pricing?
    • What if my team member can't access the Business plan features?
    • Can I test hero images with multiple slides in each variant?
    • Why does a variant not show even with 100% traffic assigned?
    • Can I export test results?
    • How do I test a new user journey that starts on different landing pages?
    • How can I preview the test without affecting analytics?
    • How do I know if my test is live?
    • How many users do I need to reach statistical significance?
    • Can I pause a test and restart later?
    • What is the difference between A/B and multivariate testing?
    • Can I run a test with just one variant?
    • Does Optibase support UTM tracking across redirects?
    • What is a 'Tested User' in Optibase?
    • How do I know when to stop a test?
    • Can I use Optibase for content localization testing?
    • Do I need to re-add conversion tracking for each new test?
Powered by GitBook
On this page
  • What is P2BB?
  • How to Interpret P2BB
  • When to Determine a Winner
  • Example
  1. What is Optibase

Probability to Be Best (P2BB)

What is P2BB?

Probability to Be Best (P2BB) is a statistical measure used in A/B testing to evaluate the likelihood that a particular variant (like Variant A vs. Variant B) is the most effective based on collected data. It is calculated using Bayesian statistics, which consider the observed results and the uncertainty inherent in testing.

How to Interpret P2BB

  • Expressed as a Percentage: P2BB is shown as a percentage. A higher percentage means a greater likelihood of that variant being the best performer in the test.

  • Direct Comparison: P2BB allows for a direct comparison between variants. For example, if Variant A has a P2BB of 70%, it suggests there's a 70% chance that Variant A is more effective than other variants in the test.

  • Informed Decision Making: Use P2BB to guide decisions on which variant to implement. A variant with a substantially higher P2BB is usually the preferred option.

When to Determine a Winner

  • P2BB is Dynamic: Keep in mind that P2BB values can change with ongoing data collection. Early in testing, the metric might be less stable and can become more reliable as more data accumulates.

  • Not Absolute: P2BB is an indicator of likelihood, not a definitive measure of success. It should be interpreted within the context of the test and alongside other business metrics.

  • Use with Other Metrics: For a comprehensive evaluation, consider P2BB in conjunction with other metrics and overall business objectives.

  • Wait for 95% Confidence: It is advisable to wait until a variant achieves a P2BB of at least 95% before declaring it as the winner. This high threshold ensures a greater level of confidence in the test results, reducing the likelihood of making a decision based on insufficient data or random fluctuations.

Example

In an A/B test comparing two website designs, if Design A reaches a P2BB of 75%, it indicates a 75% probability that Design A is better than Design B based on the current data. However, to make a more confident decision, wait until either design achieves a P2BB of 95% or more.

PreviousWhat does the custom script do?NextHow to setup Optibase

Last updated 6 months ago

๐Ÿค”
๐Ÿ‘‘