How to manually control A/B splits with box conditions

  • Created:
  • / Author: Sojib Belal

If you are planning to split site visitors into different experiences without setting up an A/B test or Multivariate test, or if you want to test a box vs “no box”, follow these simple steps.

Step 1: Go to custom JS templates

Custom JS template

Step 2: Install the “Random Percentage Rule” template

Step 3: Add a “less than” JS Variable to conditions:

DIGIOH_API.rand100sticky

And then add the percentage of visitors for which you want this box to fire.

The test is controlled by conditions that use an app feature “rand100sticky”. Each new visitor gets a random number 0-99 that is sticky in their browser. So it buckets them into 1 of potentially up to 100 test groups.

Example:

Example

So this box only shows to a random 50% of site visitors.

Tips for Analysis

A good way to analyze this is to create a Segment in GA that includes sessions with the event:

Category = Digioh
Action = Display
Label = Your box name

And also the inverse, a Segment that excludes session with that event.

Then you can run reports in GA for anything you want (e.g. bounce rate, purchase conversion rate, etc) for visitor sessions that did and did not see the box.

Questions? Comments? Let us know at contact@digioh.com