Link Search Menu Expand Document

It can be accessed under the ElasticSuite menu > AB Testing Campaign via the Optimizer Campaign entry.

Campaign creation

General configuration

ABtesting-general

Parameter Default value Description
Name   The name of the campaign
Store View Default Store View The store view for this AB Testing campaign. AB Testing campaign can only be assigned to one store view.
Description   The description of the campaign
Start Date   Start date of the campaign
End Date   End date of the campaign

AB-testing-apply-to

Parameter Default value Description
Request type   Catalog Product Search : Apply campaign only to search result listing
Catalog Product Autocomplete : Apply campaign only to autocomplete result
Category Product View : Apply campaign only to category listing

One, some or all request type can be selected

customer-segment

Parameter Default value Description
Apply to All customer segments All customer segments or Selected customer segments
Campaign can only be applied to a specific Magento customer segents

General configuration must be defined first, and then scenario A and B could be configured :

Scenario definition

Name and distribution percentage must be given to both of the scenario. If Scenario A have 60% and Scenario B 40%, it means that 60% of users will see scenario A and 40% of users will see scenario B in front. For each scenario, 3 possible choices are available :

  • Create a new boost
  • Create a new boost from existing
  • Select no boost at all (it allows to compare a scenario with boost and the other without boost)

AB-testing-scenario

Parameter Description
ID ID of the boost
Name Name of the boost
Model Constant, based on attribute value or based on behavioral data
Rule Detail of the rule used in the boost
Action 2 possible actions :
- Edit : open the popup for editing the boost
- Delete : dissociate the boost with the campaign

Publish or stop a campaign

Campaign has to be published in order to be activated when start date is reached. Otherwise it will stay on draft mode.

publish-ab-testing

If a campaign has to be stoped, a button “stop” is available. Once a campaign is stopped, it is considered completed.

stop-ab-testing

Results analytics

result-analytics

Parameter Description
Total Sessions How many sessions were exposed to the campaign
Sessions A How many sessions were exposed to scenario A
Conv. rate A Conversion rate for scenario A
Sessions B How many sessions were exposed to scenario B
Conv. rate B Conversion rate for scenario B
Significant result A campaign is significant if results are statiscally significant at 95% or more. In other words, results are significant for more than 95% of the targetted population. Statistic significance is calculated thanks to a mathematical formula that takes into account the number of session and conversion rate for scenario A and B.