A form that accepts autofill. More qualified demos.

The campaign was not the whole problem

Company was attracting visitors searching for capacity-planning and demand-forecasting software. Mobile demo requests lagged desktop, prompting proposals to reduce mobile advertising. The Director of Demand Generation wanted to know whether those visitors were a poor fit or simply struggling to contact sales.

The form failed after the buyer tried to submit

Using Daleel to compare the relevant FullSession journeys, the team isolated mobile demo-page visits with a submission attempt but no confirmation. Replays showed an error state after browser autofill: the country field appeared populated, yet its underlying selection had not been committed. The validation message appeared above the visible area.

The Web Experience Lead reproduced the failure on two mobile browser versions. Several visitors tapped Submit repeatedly, scrolled down looking for a response, and left. Captured form values were masked throughout; the investigation relied on field states and interaction behavior.

Fix the input and the recovery path

Engineering made the country field accept the autofill value, moved validation next to the affected control and brought the first error into view. The team preserved the existing qualification fields and sales routing rather than shortening the form to inflate submissions. A controlled rollout took eight working days, including browser checks and confirmation-event validation.

Measure demand quality, not just button clicks

Across a six-week test, 19,947 eligible mobile demo-page visitors entered the control and 20,083 the treatment. Raw submissions rose from 1,061 to 1,378. Using the same qualification rules, sales accepted 849 control requests and 1,080 treatment requests. Qualified-request conversion increased from 4.26% to 5.38%, or 26.3% relative. Sales accepted 80.0% of control submissions and 78.4% of treatment submissions.

After the 30-day follow-up window, 552 control prospects and 692 treatment prospects had attended a demo: 65.0% and 64.1% of accepted requests respectively. Attendance in the treatment group stayed close to the control rate.

How qualified demand was measured

Visitors were persistently randomized on their first mobile demo-page visit. Requests were attributed within seven days and deduplicated by prospect in the CRM. Qualification used the same company-size, use-case and contact-validity rules, with reviewers unaware of variant assignment. The marketing analyst joined experiment attribution to CRM outcomes outside FullSession. Closed revenue was not yet measurable because the sales cycle had not matured. Rates and relative improvement are rounded.

The operational change

Company retained the mobile campaign and added the recovered journey to its release checks. Daleel focused the analysis; FullSession sessions made the defect observable; the web team supplied the fix. The demand-generation team then adopted the accepted-request rate, rather than raw form volume, as its principal website experiment metric.