It's surprisingly easy to fool yourself with an A/B test. I've done it. You pick the metric after seeing the results, check the numbers every morning, and trust a testing platform nobody ever calibrated. None of it feels like cheating, and that's exactly the problem.
I've watched barely-significant tests get announced as big wins, and then a year later the revenue line somehow hasn't moved. Nobody sets out to be dishonest about it. The tools everyone uses are just built to declare a winner, and they never ask whether the winner is real.
So that's all this is: a small set of checks for that gap. The Validator looks at whether your platform is even telling you the truth. Lockbox locks your hypothesis in before the data lands. Reality Check deflates inflated winners before you announce them, and the Ledger quietly asks whether any of it actually reached the P&L.
They're just as much for the people on the receiving end. If test results reach you as slides from an agency or an internal team, you don't have to take the claims on faith: paste the numbers into Analyze and see what survives. Healthy skepticism doesn't require a statistics degree.
Everything runs in your browser: no accounts, no uploads. Nothing you type ever leaves your machine, so you can use them on real data without asking anyone's permission. They're free, and they'll stay free.