Seven free tools · instant test analysis

Ecommerce A/B testing you can trust.

You can make almost any test look like a winner. These tools check whether a result actually holds up, so when you call something a win, you can defend it in any room.

Analyze a test → Explore the seven tools

free · no account · runs in your browser

Bring numbers from any platform

Instant Analysis New

Been shown a winner? Drop the numbers.

Upload the results export (yours or your agency's) and get the whole verdict in one shot: is the winner real, how big is the honest lift, was the split even fair, and what to do next. Runs in your browser; nobody sees you checking.

Analyze a test →

Seven instruments.

runs in your browser · nothing you enter is uploaded
01
Platform Validator
Is your testing platform even telling the truth? Catches sample-ratio mismatch and asymmetric tracking loss before you trust a single result.
→
02
Lockbox
Pre-register the test before it runs: sample size, metric, stopping rule. Nothing left to fudge once the data lands.
→
03
Survival Curves
Two variants can tie on conversion rate and still differ in money. This finds the one that converts faster, and whether the gap is real.
→
04
Reality Check
Your winner is probably smaller than it looks. Shrinks an inflated lift back to the honest number before you announce it.
→
05
Test Receipt
A stamped, printable proof that a win was earned, not screenshotted. Tamper-evident, attach it to any readout.
→
06
Program Ledger
You announced +40% all year and revenue is flat. Reconciles claimed lift against what actually reached the P&L.
→
07
Subscriber Value
When the goal is subscriptions, analytics buries your best variant. Values a subscriber against one-off orders, even with no churn data.
→

From the blog.

a new guide most weeks →
Chart showing false alarm rates climbing with each peek at an A/B test, with a default Bayesian prior tracking the frequentist line almost exactly
Jul 19, 2026 · stopping rules · 5 min read
Does Bayesian A/B testing fix peeking?
"Our platform is Bayesian, so checking early is fine." Half right. In a simulation of 20,000 tests between two identical pages, a default flat prior called a fake winner 31.7% of the time. The prior does the protecting, not the label.
Read the guide →

Why this exists.

Most ecommerce A/B testing is governance theatre. Teams pick their metric after seeing the results, peek at significance daily, and ship winners produced by platforms nobody ever bothered to calibrate. The math was never the problem.

Each tool here guards a different failure point. The Validator checks whether your testing platform is telling you the truth. Lockbox locks your hypothesis in before any data exists. Reality Check deflates inflated winners before you announce them, and the Ledger asks the uncomfortable year-end question: did any of it actually show up in revenue?

And if tests are something that get presented to you by an agency or an internal team, treat this site as your second opinion. Paste the numbers from the deck into Analyze and see which claims survive. You don't need to be a statistician to ask the right questions; you need five minutes.

Everything runs in your browser. There are no accounts and nothing gets sent to a server, which also means you can use these on client data without asking anyone's permission.