changelog/experiments-viewshipped may 19, 2026
Most A/B testing tools hand you a table of numbers and leave you to guess whether the result is real. The new Experiments view runs your page tests through a full lifecycle and then tells you, in one plain sentence, whether the winner actually won.
Use cases
- Manage the whole run - start, monitor, and conclude an experiment from one place
- Get a straight answer - the results table states significance in plain English, not just a p-value to interpret
- Avoid false calls - see when a test simply has not gathered enough data to trust yet
Impact
You stop shipping changes on noise. Instead of squinting at conversion percentages and hoping, you get an honest verdict on whether a variant beat the control. That means the changes you roll out are the ones that genuinely move revenue, and the ones that do not get retired without drama.