HomeDev GuideAPI Reference
Dev GuideAPI ReferenceUser GuideGitHubNuGetDev CommunitySubmit a ticketLog In
GitHubNuGetDev CommunitySubmit a ticket

A/B testing

Describes A/B testing values for requesting information.

To run A/B tests that compare your existing recommendations with supplied recommendations, you can request additional A/B test information that will be supplied in the info callback. You should always include this to ensure that you know which group a user is in when in A/B testing. There are two values available:

Showing full infoShows all information available.
Showing summary infoShows only test name and A/B group.

Consider the case where an A/B test is set up to compare Optimizely recommendations with your existing recommendations. In this case, set the abTestContent attribute in the tracking request to full, and define implementations for the info and smartRecs callback functions.

The smartRecs callback is called whenever a visitor is a member of the test group set up to show Optimizely product recommendations. If the visitor is a member of the test group that is set up to show your own product recommendations, the smartRecs callback is not called.

The info callback is called regardless of the test group to which the visitor belongs. If the configs attribute contains a configuration for the test group reported in the group property, Optimizely product recommendations are shown to this visitor, and no further action needs to be taken. If the group property does not match one of the configurations, no Optimizely product recommendations are shown. Instead, you should execute JavaScript code to show your own product recommendations.

Do not rely on the group property alone to decide when your own product recommendations should be shown. For example, you write your info callback to show your product recommendations whenever the visitor's group is set to A. When the current A/B test ends and another begins which compares two competing Optimizely configurations, your site displays both Optimizely and your own product recommendations to visitors in the A group.