fieldnotes.
A little space
A little space
for big ideas.
Your notes. Your plans.
Your next good thing.
A / BA/B testing for shops and websites
More revenue from the traffic
you already have.
abdriven shows you which change on your website actually works. Build the variation, give it a fair test, and know the answer instead of guessing.
14 days free, no credit card needed. Self-hosted · first-party tracking · German and English
A clearer first step.
Illustrative example, not a live result
+16%Conversion rate in the example below
14 daystypical minimum runtime for a test
0 msflicker with anti-flicker switched on
Runs on the platform you already use
Why abdriven
Four things that make
the difference.
Revenue, not just clicks
Measure clicks, signups, carts and purchases, and compare revenue per visitor and average order value, even when the groups differ in size.
Numbers you can explain
You plan sample size and runtime before launch. abdriven uses adjusted intervals, checks the traffic split and tells you when a result actually holds.
Delivered without flicker
One small snippet, first-party tracking on your own domain and an anti-flicker option, so visitors never glimpse the old version.
Your data stays with you
abdriven runs on your own server. No third-party pixels, no visitor data handed on, and reports contain aggregated numbers only.
From idea to evidence
Three steps you
actually understand.
Build a change. Give it a fair test. Understand what the result means.
01 / ChangeSmall edit.
Small edit.
Big question.
Swap a headline, change a product image, add a section, visually, by clicking. When the idea needs more, write JavaScript and CSS. Test whole pages with a redirect experiment.
fieldnotes.
Room for
your next idea.
Text contentStart my free trial✓ saved to variation B
02 / MeasureFurther than
Further than
the button click.
Goals for clicks, signups, carts and purchases. Revenue goals compare revenue per visitor, and guardrails make sure a winner is not quietly doing damage elsewhere.
Illustrative example, not a live result
03 / DecideA result you
A result you
can defend.
Sample size, planned runtime and measurement checks sit next to the result. abdriven says plainly whether the variation won, whether it is a draw, or whether data is still missing.
Visitors per variation1,000 / 1,000
Days of testing15 / 14
Traffic split50.2 / 49.8 ✓
Adjusted interval+3.1 bis +8.9 pp
A higher rate alone is not yet a winner.
A simple starting point
Your first experiment,
in three steps.
Start with one question. Turn it into a habit.
Connect your website
Install the snippet, run the setup check, and choose the pages the experiment runs on.
Build the variation
Change the page, set the goal, check the preview, before a single visitor sees it.
Let the evidence build
Traffic is split and the plan runs. Then read the result, record it, and ask the next question.
You need a website where you are allowed to install a JavaScript snippet. Revenue tracking and some platform specifics need extra setup.
Differences
What makes abdriven different.
No enterprise buying process, no black-box statistics, no tool that fits exactly one shop system.
Agency structure is built in
Agency, clients and their websites are real levels with their own roles, not a pile of separate accounts.
AI agents can join in
Through the built-in MCP server, agents such as Claude can create experiments, read results and summarise them.
Learning does not get lost
Every result lands in the learnings library, findable, even when the team has changed a year later.
Reporting without a login
Shareable read-only report links and CSV export for everyone who only needs the result.
Check first, then launch
The setup check tells you before launch whether snippet, goals and tracking really work.
German as an equal language
Interface, notices and recommendations in German, not bolted on by machine translation.
For agencies
Many clients.
One place to learn.
Every client gets their own space with websites, experiments and results. Roles stay cleanly separated and reports go out without a login.
See the agency pageYOUR WORKSPACEExample structure
StudioAgency workspaceAgency admin
FieldnotesClient · 3 websitesClient admin
NorthlightClient · 1 websiteEditor
FAQ
Before you start.
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. You create a revenue goal and send purchase events from your website. abdriven then compares revenue per visitor alongside conversions. The snippet alone does not capture orders, that part has to be set up.
How long until I have a result?
It depends on traffic and your starting conversion rate. abdriven works out how many visitors you need before launch, and recommends at least two full weeks so weekday effects do not skew the result.
What about data protection?
abdriven runs on your own server, measures through your own domain and passes nothing to third parties. Whether you need consent depends on your setup, decide that with whoever is responsible on your side. This is not legal advice.
How do I get access?
You start with 14 days free, no credit card. We set the access up personally, there is no self-service registration yet, but somebody helps you install the snippet. You can look at the live demo at any time without access.
Make your next change an informed one
Less guesswork.
More now we know.
One question is a good place to start.