Documentation

DocsA/B Testing

A/B Testing

Never guess what resonates with your audience. AlgorithmX natively supports comprehensive A/B Testing across campaigns (Email, Web Push) and product recommendations.

A/B testing allows you to send two or more variations of a campaign to a subset of your audience, analyze which performs better, and deploy the winning version to the rest of the audience.

1. What Can You Test?

You have complete flexibility to test multiple elements of your marketing strategy:

  1. Different Messages: Test subject lines, preheaders, or push notification titles to see which drives higher Open Rates.
  2. Different Offers: Test a "20% Off" discount against a "Free Shipping" offer to measure which generates more revenue.
  3. Different Designs: Compare two entirely different email layouts (e.g., text-heavy vs. image-heavy) created via the Design Editor or HTML Editor.
  4. Different Audience Segments: Test the exact same message on two different behavioral segments to evaluate audience responsiveness.

2. Step-by-Step Guide

Step 1: Create a Campaign Variation

  1. While building an Email or Web campaign, navigate to the Campaign Variation step.
  2. Click Add Variation. This creates a parallel track for your campaign (e.g., Variation A and Variation B).

Step 2: Configure the Differences

  1. Modify the specific variable you want to test in Variation B.
  2. Best Practice: Only change one variable at a time (e.g., only change the subject line). If you change the subject line AND the design, you won't know which change caused the improvement.

Step 3: Define the Split

  1. Specify how the audience should be divided. For example, send Variation A to 50% of the segment, and Variation B to the other 50%.
  2. Alternatively, send a test to 20% of the audience (10% A / 10% B), wait 4 hours, and automatically send the winning variation to the remaining 80%.

Step 4: Analyze Results

Once the campaign is launched, navigate to the Analytics dashboard. AlgorithmX will clearly display the performance (Open Rates, Click Rates, Conversions) side-by-side, allowing your team to confidently choose the best-performing option.