A/B Testing
Business Feature
What is A/B Testing?
With A/B Testing, you compare two variants of an action (e.g. two different bundles) and find out which one converts better. Customers are automatically split 50/50 between the variants.
Supported Modules (1.0)
Module | Description |
|---|---|
Bundles | Compare two bundle configurations (e.g. 10% discount vs. fixed price) |
Upsell Popups | Compare two popup variants (e.g. different products or headlines) |
Additional modules will be added in future updates.
Create an A/B Test
Marketing > Marketing Suite > Analytics > Tab "A/B Testing"
Click on "+ Create New Test"
Fill in the fields:
Module Type: Bundle or Upsell Popup
Select Variant A: Choose an existing rule (dropdown shows name + discount info)
Select Variant B: Choose a different rule of the same type
Test Name: Generated automatically (e.g. "Bundle: Ticket Bundle vs. Fixed Price Bundle")
Primary Metric: How success is measured (Orders, Revenue, Conversion)
Click on "Create Test"
Important: Variant A and B must be different rules.
Start Test
After creation, the test has the status Draft. Click on "Start" to activate the test.
When starting:
A snapshot of the configuration of both rules is saved
The test becomes active on the storefront
Customers will now only see one of the two variants
Limitation: Only one test per module type can be active at a time. A second bundle test can only be started after the first one is stopped.
How does assignment work?
Each visitor is assigned to a variant via cookie:
Logged-in customers: Deterministic assignment based on customer ID (same customer always sees the same variant)
Guests: Random 50/50 assignment
Cookie duration: 30 days
The cookie is called "fgits_ab_{testId}" and contains "A" or "B".
Privacy Notice (GDPR)
The A/B Testing cookie is not a technically necessary cookie. If you use A/B Testing, make sure to add the cookie to your cookie banner or consent manager under the category "Statistics" or "Marketing". The cookie name is fgits_ab_*, duration 30 days, purpose: Assignment of an A/B test variant.
Metrics
Per variant, the following are tracked:
Metric | Description |
|---|---|
Impressions | How many times the variant was displayed |
Triggers | How many times the action was triggered |
Acceptances | How many times the customer accepted the action |
Orders | Orders with this variant |
Revenue | Influenced order revenue |
Discount Costs | Discount granted |
Evaluate Results
In the test detail view, you see a comparison table:
Metric | Variant A | Variant B |
|---|---|---|
Impressions | 142 | 138 |
Orders | 12 | 8 |
Revenue | 480 EUR | 320 EUR |
Trend Display
With sufficient data (at least 5 orders per variant): Currently leading: Variant A
With insufficient data: Not enough data yet for a reliable trend
Important: The display shows a trend, not statistical significance. For reliable results, we recommend at least 50+ orders per variant.
Choose Winner
Open the test detail view
Check the metrics
Click on "Choose Variant A as Winner" or "Choose Variant B as Winner"
The test is ended and the winner is marked
Status Transitions
From | Action | To |
|---|---|---|
Draft | Start | Active |
Active | Pause | Paused |
Paused | Resume | Active |
Active/Paused | Stop | Completed |
Active/Paused | Choose Winner | Completed |
Planned Enhancements
Automatic winner detection
Time-scheduled test runs with automatic end
Result significance (e.g. sufficient data available)
Tests for different customer groups