🚀 2025-22-3 (Unscheduled)

Release date: April 17, 2025

TL;DR : This unscheduled release includes improvements to the A/B testing result reports.

Subscribe to Bloomreach Discovery release email notifications by filling out this form.

Improvements


1. Consistent A/B test result metrics

A/B test result metrics are easier to understand and use in the dashboard.

What improved?

  • All metrics displayed in test results have tooltips that show the metric's meaning or calculation.

  • The definition of “Visits” has been updated. Previously, the “Visits” value matched the number of browse sessions. Now, visits are calculated separately—a single visit may include multiple browse sessions.

  • Confidence calculations are more accurate. Confidence is not calculated if the test has run for less than 2 days or if there are fewer than 1000 visits. In this case, the default confidence is 0.

  • Test bucket names can now have another bucket's name as a prefix without conflicts.

Understanding the difference between Browse sessions and Visits

To clarify this change, let’s look at a quick example: A shopper visits your site and performs the following actions -

  1. Searches for “dress”.
  2. Then searches “jacket”.
  3. Then clicks on the “Accessories” category.

In this case, each search or category click starts a new browse session. Since all of this activity happens during a single continuous visit, it counts as one visit.
So, you’d see:

  • Browse sessions = 3
  • Visits = 1

Note: This definition of “browse session” is specific to Bloomreach and may differ from generic web analytics tools.

Why is this valuable?

These updates allow you to interpret the A/B test results faster and offer additional insights while making decisions.

Visit the documentation article to learn more.