A/B Testing Explained

A/B Testing - shows 2 tests with equal users one results in more conversions = winner
Image referenced from https://www.invespcro.com/

How A/B Testing Works

A/B testing, also known as split testing, is a method of comparing two versions of a webpage, app screen, ad, email, or other asset to determine which one performs better. The goal is to improve user experience, engagement, and conversions. Image source


  • Identify a goal: You need to define what you want to optimize for. This could be clicks, conversions, time on page, etc.
  • Determine your variables: Decide what element you want to test. This could be the page layout, headline, call-to-action, image, etc. Create a version A and B.
  • Split traffic: Send a portion of your website or app traffic to version A and a portion to version B. 50/50 splits are common.
  • Measure results: Use analytics to track how each version scores for your defined goal. Let the test run until you have statistical significance.
  • Pick a winner: The version that provides better results against your goal is the winner. Move forward with this version.

“A/B testing is the cornerstone of data-driven decision-making. It allows you to objectively compare different options and identify which ones perform best for your target audience.”

Eric Ries, Author of The Lean Startup

Goals of A/B Testing

  • Increase conversion rates and revenue.
  • Improve user experience and engagement.
  • Test design changes like colours, copy, layouts.
  • Identify the most effective call-to-action.
  • Reduce bounce rates.
  • Determine optimal page elements and messaging.
  • Test if a change reduces friction for the user to take an action or make a decision.

Criteria for Effective A/B Tests

  1. Have a clear goal you are optimizing for.
  2. Allow tests to run long enough to achieve statistical significance.
  3. Change only one variable at a time.
  4. Use proper analytics tracking.
  5. Make relevant test variations to learn something.
  6. Ensure enough traffic volume to both versions.
  7. Analyse results correctly to pick the right winner.

Following these best practices will allow you to conduct high-quality A/B tests that yield reliable results.

Hazel Rigby – SunLife Webmaster