The demo opens Probant on a fictional store.

Maison Lune doesn’t exist. Its tests, visitors, and orders are demo data, analyzed by the same engine as your future tests.

Explore the demo Visiter la boutique

No sign-up
Read-only
Data from a fictional store

8 tests, 5 different statuses.

Differences, ranges, and progress are computed by the engine from the demo store’s visitors and orders. Each row opens the matching report in the demo.

Maison Lune ExperimentsDemo store

Tests of the Maison Lune demo store, with their status, progress, and measured difference
Test	Status	Difference and range
Free shipping progress barCart · Revenue per visitor	Conclusive. Proven gain61,240 visitors	+6.0%+2.2% to +10.0%
One-page checkoutCheckout · Conversion	Sample ratio mismatch.19,800 visitors	Results hidden until the split is diagnosed
Simplified mobile menuNavigation · Revenue per visitor	Not conclusive yet…day 9 of 13	+1.9%-3.7% to +7.8%
Pay-in-3 option highlightedProduct page · Revenue per visitor	Not conclusive yet…day 6 of 17	+3.5%-3.4% to +11.0%
Customer reviews at the top of the pageProduct page · Revenue per visitor	No detectable effect.64,100 visitors	-0.4%-4.0% to +3.3%
“You may also like” recommendationsProduct page · Revenue per visitor	Stopped by a guardrail.stopped 3 h ago	Returns +1.2 pts
Orbe product page: title, photo, and badgeProduct page · Add to cart	Conclusive. Proven gain80,512 visitors	+26.6%+13.9% to +41.0%
Sort by popularityCategory page · Product click	Not conclusive yet…day 3 of 14	+1.5%-2.8% to +6.1%

What each status means.

The word says where the test stands, and the final dot gives the result: green for a proven gain, orange for a proven loss, amber for an alert. So “Conclusive.” is read together with its dot, as the loss row shows.

Conclusive. Proven gain: Proven gain. The whole range is on the right side at a planned checkpoint. The report gives the expected revenue and margin in euros, each with its range. In the demo: “Free shipping progress bar”.
Sample ratio mismatch.: Visitors aren’t split as planned. Results stay hidden while the cause is diagnosed. In the demo: “One-page checkout”.
Not conclusive yet…: The test runs until its next planned checkpoint. The day’s difference is shown, with no conclusion. In the demo: “Simplified mobile menu”.
No detectable effect.: The planned sample size is reached and the range spans zero. The report says what it would take to decide. In the demo: “Customer reviews at the top of the page”.
Stopped by a guardrail.: A guardrail metric went past its limit, and the test stopped on its own. In the demo: ““You may also like” recommendations”.
Conclusive. Proven loss: Proven loss. The whole range is on the wrong side: the variant does worse than the original. In the demo library: “Strikethrough prices in red”.

7 rollouts, from the first step to rollback.

Winners roll out step by step. Guardrails are monitored at each step, and a proven drop sends everyone back to the original. The engine replays dates and steps from each day’s data.

Free shipping progress barCart · Winner · decision pending
Not started yet
1. 25%: upcoming
2. 50%: upcoming
3. 100%: upcoming
Free shipping from €60Cart · Winner in October
50% step in progress
1. 25%: passedOct 5 → Oct 7
2. 50%: in progresssince Oct 8
3. 100%: upcomingdate to come
Not enough data to move to 100%. The range for JavaScript errors still overlaps its limit, and the estimate for JavaScript errors is already on the wrong side, so more data won’t bring it closer.
Free returns highlightedProduct page · Winner in September
Rolled out to 100%
1. 25%: passedSep 14 → Sep 16
2. 50%: passedSep 17 → Sep 27
3. 100%: passedsince Sep 28
From 25% to 100% in 14 days, with no rollback.
Simplified filters on mobileNavigation · Winner in August
Rolled out to 100%
1. 10%: passedAug 17 → Aug 20
2. 25%: passedAug 21 → Aug 23
3. 50%: passedAug 24 → Aug 30
4. 100%: passedsince Aug 31
From 10% to 100% in 14 days, with no rollback.
Shipping costs shown on the product pageProduct page · Winner in July
Rolled out to 100%
1. 10%: passedJul 13 → Jul 15
2. 25%: passedJul 16 → Jul 18
3. 50%: passedJul 19 → Jul 26
4. 100%: passedsince Jul 27
From 10% to 100% in 14 days, with no rollback.
Guest checkoutCheckout · Winner in June
Rolled out to 100%
1. 25%: passedJun 22 → Jun 24
2. 50%: passedJun 25 → Jun 29
3. 100%: passedsince Jun 30
From 25% to 100% in 8 days, with no rollback.
Simplified order summaryCart · Winner in March
Rolled out to 100%
1. 25%: passedMar 16 → Mar 19
2. 50%: passedMar 20 → Mar 25
3. 100%: passedsince Mar 26
From 25% to 100% in 10 days, with no rollback.

Look at everything. Nothing is saved.

The demo opens without an account and stays read-only. Once you log in, you land in your workspace, and the menu takes you back to the demo.

OverviewToday’s summary, the impact measured against the holdout, the decisions to make, and running tests.Open the overview A winning reportThe one-sentence verdict, the measured lift and the lift corrected for the winner’s curse, expected revenue and margin in euros, checks, and guardrails.Open the report The variant editorA draft test on the demo product page: changes, automatic checks, and the assistant, which answers with prepared replies.Open the editor The learnings libraryThe store’s 37 finished tests and 12 tests imported from AB Tasty, their effects on one scale, and lessons by theme.Open the library RolloutsRollout steps, guardrails checked at each step, rollback, and the feature flags tab.Open rollouts A test planPre-launch checks, plus the duration and checkpoints calculated before the first visitor.Open the plan

To launch a real test, install the script, or invite your team, you need a workspace: create your workspace.

Conclusive Or not yet…

Create your workspace, or explore the demo first with data from a fictional store.

Create your workspace Explore the demo

Open demo, no sign-up
Free to start, no overage charges
Database and compute in Amsterdam, with Scaleway