Modal Form vs In-Page Form A/B Test Example

Heatmaps where few visitors scroll down to the form, session recordings of people passing over it, or visitors who leave the form halfway to go back to the page's explanation show that the form gets overlooked or that visitors need the information behind it while they fill it in. The expected behavior is that the better presentation draws more intentional interaction with the form and more completed submissions. Launch the modal arm only with a visible close button, keyboard dismissal and a trigger rule written down in advance; if the problem lies in the fields or validation, fix that first.

Form completion rate

Assign eligible visitors before they see either presentation and compare successful submissions.

Also track

• Form start rateMeasure intentional first interaction with the form the same way in both arms, not modal opens.
• Data loss after modal dismissalThis diagnostic exists only in the modal arm; do not compare it directly across arms.
• Page exit rateCompare page exits among visitors assigned to each form presentation.

• PresentationShow the same form inline in the page or in a modal.
• Keep constantKeep fields, validation and support information the same; define when the modal opens before launch.
• Data persistenceCheck whether entered information is preserved when the modal is closed and reopened.
• Browser behaviorVerify the browser back button behaves as expected in both presentations.
• Accessibility and deviceCheck keyboard dismissal, focus containment and usability on small screens.

1. 1Compare form completion rate across all eligible assigned visitors, including those who never open the modal. Data loss after modal dismissal is a diagnostic that exists only in the modal arm; a modal that silently wipes entered data on an outside click is not a winner, however many submissions it collects.
2. 2Count a form start as the first intentional interaction with the form, not as a modal open. Do not weigh a modal squeezed into half of a mobile screen against a desktop result; read each device in its own layout. Treat the outcome as specific to this form on this page.
3. 3Backdrop darkness and the trigger rule must stay fixed for the whole test; if either changed, the result no longer measures presentation alone. Because visitors are assigned before they see either presentation, confirm the arms match the planned split, then compare the two presentations' form completion rates in the A/B test significance calculator. Should the gap remain unclear, choose by task length and how much context the form needs.