Not just A against B
Besides the control group you can add several variants, each with its own layout and traffic share. The sum of shares is validated so nothing exceeds one hundred percent.
Test instead of guessing.
Run two shopping experiences against each other, split the traffic as you like and measure which variant actually sells - not which one wins the meeting. With significance testing, sample size estimation and one click to publish the winner.
What the plugin doesNew
Variants head to head, numbers decide.
You create an experiment, pick the original shopping experience and one or more variants. Visitors are assigned at random and stay with their variant. The plugin measures, calculates significance and tells you when the result is solid.
Visitors keep their variant
Whoever has seen variant B sees it again on the next visit. Otherwise experiences mix and the measurement becomes worthless.
When it counts
The confidence level is configurable. A difference only counts as solid once it is reached and the minimum sample per variant is in - not after a three order lead.
Know the duration upfront
From the expected conversion rate and the smallest detectable difference the plugin estimates the sample needed per variant. You know before starting whether your traffic supports the test.
Conversion and revenue per visitor
Measured are conversion rate, revenue per visitor and the path through the funnel - not just where more clicks happened.
One click, with a way back
The winner can be published; the original page is overwritten but a rollback stays available. On request the winner goes live automatically when the test ends.
What it is for
Opinions are not data.
Home pages and category headers get discussed a lot and measured rarely. A test ends the debate with numbers instead of hierarchy.
Compare the opener
Two versions of the home page against each other - different hero, different order of elements, different intro copy. After a few weeks you know which one carries.
Structure under test
Description above or below, filters open or closed, tiles large or small. Small changes with measurable effect on revenue per visitor.
Step by step instead of big bang
A new layout does not have to hit everyone on day one. Twenty percent first, measure, then decide - that takes the risk out of a relaunch.
Control
Measure cleanly, without side effects.
A test must not disturb the shop. That is why you can define exactly who takes part and what gets counted.
Only where it matters
An experiment can be limited to one sales channel. Without a selection it applies to all - useful with several brands in one system.
Crawlers do not distort
Bots and crawlers can be excluded from assignment. Otherwise search engines count as visitors but never buy.
No premature verdict
Per variant you can require a minimum number of impressions before any result is shown at all.
Stop at any time
Experiments can be started, paused and completed. Completing determines the winner; publishing only happens on your command.
No orphaned clones
Variants create copies of the shopping experience. A cleanup function removes clones that no longer belong to an experiment.
Back to the original
Even after publishing, the original page can be restored. No decision is ever final.
Technical details
Clean Shopware integration.
Variants are real shopping experiences and assignment happens server side. Installed by upload in the admin, without SSH.
Compatibility
Server side assignment
The variant is determined when the page is requested, not swapped in by a script in the browser. There is no flicker, and measurement does not depend on a script running.
Reporting of your own
Impressions, orders and revenue are stored and evaluated in the shop. There is no connection to an external analytics service and no data sharing.
Variants as copies
Every variant is a shopping experience of its own and stays editable in the editor. After the test, copies that are no longer needed can be removed in one go.
Frequently asked
Frequently asked questions about A/B testing
Significance, effort and what actually gets measured.
How do I run A/B tests in Shopware 6?
You create an experiment with an original shopping experience and one or more variants and split the traffic. Visitors are assigned at random and stay with their variant so the measurement stays clean.
How long does an A/B test need to run?
That depends on your traffic and the difference you want to prove. From the expected conversion rate and the smallest detectable difference the plugin estimates the sample size per variant before you start.
What does A/B testing cost?
Buying directly costs 19.99 euros net per month, 99.99 euros net per half year or 179.99 euros net per year. Paying yearly saves around 25 percent. All prices are net and exclude VAT.
What is measured, only clicks?
No. Measured are conversion rate, revenue per visitor and the path through the funnel. That shows which variant actually sells, not just where more clicks happened.
Can I undo a decision?
Yes. Publishing overwrites the original page and a rollback restores it. Variants no longer needed can then be removed in one go.
Availability
Through the store or directly from us.
Two routes, the same feature set - pick whichever suits your setup.
Available now
Monthly19.99 €
Half-yearly99.99 €17 % cheaper
Best priceYearly179.99 €25 % cheaper
All prices net, plus VAT.
You buy the licence directly from us and receive an invoice from us. Installation by upload in the admin without SSH, updates through our licence platform, support straight from the developer.
In preparation
The plugin will be listed in the official Shopware Community Store. Installation and automatic updates then run through the plugin manager in the backend, billing through your Shopware account. We will link it here as soon as it is available.