Split traffic between published versions of a form and let the numbers pick the winner. Built into the form editor — no extra tool, no scripts.
Each variant is a published version of your form, so you can compare new wording, a different layout, or a reordered flow against what's live today. Mark one variant as the control.
Split 50/50, 80/20, or across more than two variants. Weights always add up to 100%, and every submission records which variant it came from.
Every variant is compared against the control with a two-proportion z-test and 95% confidence intervals, so you can tell a real difference from noise.
When a variant beats the control with statistical significance, it's flagged as the winner. A sample-size estimate shows how many more responses you need before the result is reliable.
Run one experiment per form at a time. Pause it without losing data, and complete it once you have your answer.
Testing a form usually means wiring in a separate optimization tool with scripts. In formcel, experiments live in the form's settings, and the results sit right next to them.
Run your first experiment on a form you already have. Publish a new version, split the traffic, and watch the results come in.