Know what pays off, without being a statistician.
Probant calculates the test length before launch, monitors guardrails while the test runs, and concludes in one sentence with a figure in euros: revenue and margin, each with its range.
Free shipping progress bar
Conclusive. Proven gainShowing progress toward free shipping: revenue per visitor +6.0% measured, +4.9% expected once corrected for the winner’s curse. The effect on margin isn’t established yet: +3.0% measured, range from -0.7% to +6.7%.
Expected revenue, corrected for the winner’s curse+€125,700 per month
Maison Lune, a fictional demo store · report computed by the engine
A misread test costs more than a test you never ran.
The most common problems when you test your own store between packing orders.
- A winner that vanishesThe lift looks clear after a few days, then fades. Deciding by checking the test every day produces false winners.
- Revenue up, margin downA discount or free shipping can inflate revenue and cost more than it brings in.
- No time for statisticsP-value, power, interval: the jargon gets in the way before you reach the result.
- No developer availableEvery variant waits its turn in the tech team’s queue, or an agency’s.
- Modest trafficNot every idea can be measured in a month. Better to know before you launch.
Probant does the math and writes the conclusion.
Each feature below is illustrated with the demo store.
Maison Lune, a fictional demo store
The verdict counts margin; the variant is built on the page.
Revenue and margin, side by side.
“Free shipping progress bar” raises revenue per visitor by +6.0% (from +2.2% to +10.0%) and has no established effect on margin per visitor (+3.0% measured, from -0.7% to +6.7%). Free shipping costs the store €23,800 a month, already deducted from margin.
A discount that raises revenue and lowers margin shows up in the same report, before any rollout.
Free shipping progress bar
- Revenue per visitorprimary+6.0%+2.2% to +10.0%
- Margin per visitorafter product costs, discounts, and shipping+3.0%-0.7% to +6.7%
- Orders with free shippingexplains the gap between revenue and margin41.0% → 57.0%+16 points
Build the variant on your own page.
Click the element you want to change, or describe the change in one sentence to the assistant. Probant checks the variant before any visitor sees it: weight added to the page, stable selectors, text contrast.
Assistant
Changes
Automatic checks
From idea to rollout, no spreadsheet.
How a test runs on your store, and what the demo store shows at each step.
- Choose
Take the top-ranked idea.
Ideas are ranked by expected gain and effort. The ones your traffic couldn’t measure are flagged.
Demo storeFree shipping from €50: expected effect +2.5% based on 4 shipping tests, 14 days to conclude. - Build
Build the variant in the editor.
On your live page, by clicking the element to change or describing the change to the assistant.
- Launch
Launch with a locked plan.
Test length, target effect, and checkpoint days are set at launch, so the decision is made at the planned checkpoints, not whenever it suits someone.
Demo storeFree shipping progress bar: 18 days, checkpoints on days 6, 12, and 18. - Let it run
No daily monitoring.
Guardrails are read at every check. A proven drop stops the test and notifies you.
- Decide
Read the conclusion.
One sentence, one range, and a monthly amount corrected for the winner’s curse. The details stay available if you want to check them.
Demo store+€125,700 in expected revenue a month, range +€41,600 to +€209,800. - Roll out
Put the winner live.
In steps, with the same guardrails. At the end, the variant is served to everyone without changing your site.
Demo storeFree returns highlighted: rolled out to all visitors on Sep 28.
What holds for every store.
Frequently asked questions.
Do I need to know statistics?
No. Probant calculates the test plan before launch, sets the checkpoint days, and writes the conclusion in one sentence. The detailed calculations stay available if you want to check them.
Do I have enough traffic to test?
Probant tells you before launch. In the demo store, the cart gets 3,400 visitors a day: in 30 days, a test there can detect a 3.5% effect on orders, or 1.7% on reaching checkout, which happens more often. A smaller effect takes longer, and Probant tells you so before launch.
Do I need a developer?
Only to install: someone has to add the script to your site’s <head>. After that, you create variants in the visual editor, by clicking on the page or by describing the change to the assistant.
How long does a test run?
As long as it takes to detect the effect you’re aiming for, calculated before launch. The demo’s “Free shipping progress bar” test was planned for 18 days and 58,900 visitors.
What if a test hurts sales?
The guardrails set in the test plan (return rate, mobile speed, revenue per visitor) are monitored during the test and during the rollout. A proven drop stops the test or sends visitors back to the original.
Other use cases
The same engine, for other site areas and other teams. The demo opens on the screen this page describes.
- Open demo, no sign-up
- Free to start, no overage charges
- Database and compute in Amsterdam, with Scaleway