Visual editor
Change text, images and
buttons by clicking.
A headline, an image, a button or a whole section.
How the editor worksA/B testing for shops and websites
More revenue from
the traffic you already have.
Test changes to your website and see what really brings more revenue.
- Element picked · button text
- Saved to variation B
- Preview link created
fieldnotes.
Room for
your next idea.
Button textGet startedStart my free trial
- Test started · 50 / 50
- Checkout guardrail holding
- 1,000 visitors per variation reached
Shipping note in the cart
RunningVisitors per version1,300/ 3,000
Original10.0%Conversion rate
With shipping note12.2%Conversion rate
Days of testing9/ 14
OriginalWith shipping noteConversion rate
Day 1Day 7Day 14
- Variation B wins · +60%
- Chance to win above 95%
- Report link created
Clearer add-to-cart button
A clearer label helps visitors take the next step.
CompletedSetupResults
Result
Clearer button performed better
The results support using this version on your website.
Improvement confirmed
Conversion rate
Original10.0%
Clearer button16.0%
60.0% more conversions per visitor.
Measurement
Planned checks passed
Visitors per version1,000 / 1,000
Days of testing15.0 / 14
Traffic split checked.
All goalsThe main goal decides
VariationVisitors
Add to cartRate · conversionsMain goal
CheckoutRate · conversionsGuardrail
PurchasePer visitor · total
AOriginal1,000
Baseline10.0% · 100
Baseline33.4% · 334
Baseline€1.00 · €1,000.00
BClearer button1,000
+60.0%16.0% · 160>95% chance to win
0.0%33.4% · 334≈50% chance to win
+68.0%€1.68 · €1,680.00
- Winner served to everyone
- Extra revenue from winners: +€1,385
- Learnings saved
Overview
+ New experimentExtra revenue from winners+€1,385From 1 winning test, since it went live
Extra revenue per month+€2,760Estimate, if the effects hold
Running tests1Nothing else waiting
Win rate100%1 of 1 completed tests won
Clearer add-to-cart button
Serve the winner to everyone, or build it into the website.
Original50 %0 %Clearer button50 %100 %
Runs on the platform you already use
Change text, images and
Platforms
For Shopify, WooCommerce,
Shopware, Webflow and Magento.
One line of code. The setup check makes sure everything arrives before launch.
Shopify, WooCommerce and moreStatistics
Results with win probability
and guardrails.
- Chance to win and adjusted intervals
- Guardrails, so a winner breaks nothing else
- Revenue per visitor, not just clicks
abdriven dashboard
All tests and results
in one place.
Drafts, running tests and winners. If you like, AI agents take care of the routine.
Extra revenue, running tests and everything that needs your decision right now, at a glance.
Overview
+ New experimentExtra revenue from winners+€1,385
Win rate100%1 of 1 completed tests won
Running tests1Nothing else waiting
Needs your input1
Clearer add-to-cart button
Open resultsServe the winner to everyone, or build it into the website.
Every goal, every variation, with chance to win and guardrails. The main goal decides, the rest protects.
Clearer add-to-cart button
CompletedAll goalsThe main goal decides
VariationVisitors
Add to cartRate · conversionsMain goal
CheckoutRate · conversionsGuardrail
PurchasePer visitor · total
AOriginal1,000
Baseline10.0% · 100
Baseline33.4% · 334
Baseline€1.00 · €1,000.00
BClearer button1,000
+60.0%16.0% · 160>95% chance to win
0.0%33.4% · 334≈50% chance to win
+68.0%€1.68 · €1,680.00
Claude, Cursor or ChatGPT create drafts and read results through the MCP server, within the permissions you grant.
AI agent over MCP
Within your permissionsTest a shipping note in the cart and plan how long the test has to run.
create_experimentDraft created: shipping note in the cartplan_sample_size3,900 visitors per variation, about 14 dayscreate_preview_linkPreview link ready for both variations
The draft is ready. You can start it once you have checked the preview.
Industries
Goals and tests
for your industry.
More clicks do not help if baskets get smaller.
More on E-commerceMore sign-ups are not always better.
More on SaaSB2B sites have few visitors, but every enquiry counts.
More on B2B and leadsWeekdays, holidays and prices swing.
More on Travel and bookingEvery subscription hurdle costs reach.
More on Media and publishersFor developers and agents
Small snippet, own domain,
API and MCP.
A small snippet on your own domain, a setup check before launch, and API tokens with exactly the rights they need.
Read the documentationDeploy agents with our MCP server
Connect Claude, Cursor or ChatGPT to abdriven and let them create tests and read results, within the permissions you grant.
FAQ
Questions about abdriven
What can I test with abdriven?
Headlines, images, buttons and whole sections in the visual editor. More involved ideas with JavaScript and CSS. And completely different pages as a redirect experiment.
Will it work on my platform?
Anywhere you are allowed to install a JavaScript snippet. Shopify, WooCommerce, Shopware, Webflow and Magento have their own guides including revenue tracking.
Can I measure purchases and revenue?
Yes. Create a revenue goal and report purchases from the order confirmation. abdriven then compares revenue per visitor.
How long until I have a result?
It depends on your traffic. abdriven works out beforehand how many visitors you need. Plan at least two full weeks.
What about data protection?
abdriven runs on your server and passes nothing to third parties. Whether you need consent depends on your setup. This is not legal advice.
How do I get access?
Through the “Request access” form. We get back to you and set everything up with you.