Replacing written application answers with multiple choice raised submissions by 56.7% in a webinar-to-call funnel, alongside sales-team reports of poorer leads and attendance. Those reports did not establish a fall in purchases per assigned visitor. In a separate booking-and-application test, completion rose from 13.94% to 17.56%, rounded to a 26% relative lift. Start with duplicated data entry, then test qualification questions against purchases or qualified attended calls per assigned visitor.
What job does each question do?
For every field, name the decision its answer changes: eligibility, routing, call preparation, consent, scheduling or follow-up. If nobody uses an answer and it does not satisfy a real requirement, the field is likely waste. If the answer predicts a valuable downstream outcome, removing it can make the form easier while making the funnel less useful.
| Field purpose | Keep when | Reduce effort by |
|---|---|---|
| Objective fact or routing | The answer changes eligibility or destination | Use complete, mutually exclusive choices |
| Context for a sales conversation | The substance improves preparation or fit assessment | Ask one focused open question |
| Readiness or timing | The answer changes prioritization or follow-up | Use neutral wording and a clear time range |
| Known contact or booking data | The system already has a reliable value | Prefill or integrate instead of asking twice |
| No downstream use | No owner or decision depends on it | Remove the field |
Case 1: open answers versus multiple choice
The application sat between a webinar and a call for a high-ticket program. The sales team used the written answers to decide whether to allocate a call slot. We replaced those answers with multiple choice, working with the client to rewrite questions that did not translate directly.
We split traffic evenly for ten days. Submissions rose 56.7% relative to control. The sales team then reported unqualified applicants, no-shows and a decline in application-to-purchase conversion.
Because we rewrote questions as well as changing the answer format, this tested the complete form. The later reports do not establish its effect on purchases per assigned visitor.
Application-to-purchase conversion uses a population the treatment itself changes: the people who submitted. To decide whether the extra applications are worth pursuing, compare purchases or qualified attended calls per originally assigned visitor, with the same follow-up window.
As an arithmetic illustration, with 56.7% more applications, application-to-purchase conversion could fall to about 63.82% of its former level before purchases per assigned visitor fell, assuming comparable populations and follow-up. This is not a measured client outcome.
Case 2: removing an urgency question
In a separate test, we removed one question about readiness to act. We ran it for five days with 4,210 participants, using form submission as the primary metric and attendance, cancellations and applicant mix as secondary checks.
Submissions increased by 4%, while show-up fell by 4%; fewer applicants were in the high-income category, and both client-initiated and applicant-initiated cancellations rose. The published percentages do not distinguish relative changes from percentage points, and the historical call-completion and cancellation display did not mark those differences statistically significant.
Our hypothesis was that the question prompted useful reflection before booking. It remains a hypothesis: income alone is not a definition of fit, and these results do not establish that effort creates commitment.
If readiness matters, ask it neutrally—for example, “When would you be ready to begin if the program is a fit?”—and connect each response to an actual follow-up decision. Avoid manufactured pressure or vague emotional language that the team cannot interpret consistently.
Case 3: combining booking and application
After an on-demand webinar lasting about an hour, visitors chose a call time in OnceHub and then had to complete a separate Wufoo application. Both systems asked for personal details. If the second form was not completed within two hours, the call could be cancelled.
| Stage | Control | Combined treatment |
|---|---|---|
| Choose a time | OnceHub booking | OnceHub booking |
| Supply application answers | Redirect to mandatory Wufoo form | Questions built into OnceHub after time selection |
| Contact details | Requested in both systems | Collected once |
| Finish the application | Second form due within two hours to keep the call | Required answers collected in the booking flow |
We moved the application questions into OnceHub and removed repeated personal fields. This changed the handoff and cancellation behavior together; the test evaluates that package, rather than the number of screens alone.
Across 1,409 participants over five days, complete-application rates rose from 13.94% to 17.56%: a 3.62-percentage-point increase and a 25.97% relative lift, rounded to 26%. Significance remains unconfirmed; a follow-up should check attended calls and purchases.
| Measure | Control | Combined treatment |
|---|---|---|
| Complete-application rate | 13.94% | 17.56% |
Keep the scheduling and qualification rules intact
Create a field map before merging systems: identify the source of truth for contact details, qualification answers, consent, selected time, timezone and routing fields. Remove true duplicates, not information required for scheduling, compliance or qualification. Confirm that validation errors retain the chosen slot and entered answers.
Test expired slots, double submissions, API failures, incomplete required fields, cancellation timing and CRM delivery. Record a completed application only after all required data is accepted. A calendar selection starts the booking; it does not complete the application.
Keep separate stages when people need to save progress, gather information or answer conditional questions, or when qualification should happen before scarce calendar slots are reserved. Integration should remove repetition while preserving those decisions.
What the three tests change about form design
| Change | Population and duration | Observed result | Decision and uncertainty |
|---|---|---|---|
| Written answers → multiple choice | Webinar-to-call applicants; ten days, even traffic split | +56.7% submissions; reported lead-quality and attendance concerns | Retest questions used to assess fit; the change in purchases per assigned visitor was not established |
| Remove readiness question | 4,210 participants; five days | +4% submissions; reported −4% show-up and more cancellations | Compare attended calls; call-completion and cancellation differences were not marked significant, and the 4% units were not specified |
| Merge booking and application | 1,409 participants; five days | 13.94% → 17.56% completion | Remove repeated entry and the second-system handoff; significance remains unconfirmed |
Use structured choices for facts and routing, and a focused written answer when the substance helps someone prepare for a call. A question earns its place through the decision its answer changes, not through the effort it demands.
Connect submissions to sales outcomes
Bind experiment assignment to the CRM without exposing personal data in analysis. Report assigned visitors, completed applications, sales-accepted leads, booked calls, attended calls, purchases, refunds and net revenue by arm over the same window. Define qualification before launch and record missing fields, cancellations and follow-up time consistently.
If conversion signals are sent to ad platforms, do not assume the easiest form completion is the outcome the bidding system should optimize. Use a reviewed qualified-lead or purchase signal only when consent, identity matching and value definitions are sound.
Choose the qualification criteria with the sales team, then record when a lead meets them in the CRM. Keep that event separate from an application submission. To send it from PostHog, use the Meta destination guide or the Google Ads destination guide; the useful signal is the outcome the business values, not simply the easiest action to increase.
Where to test next
Start with repeated fields and broken handoffs, as in the merged booking flow. At acquisition, a form has a different job: our pre-lander test made the questions shorter and more conversational, then checked applications after the webinar. Earlier in the journey, our webinar timing tests compare delays between registration and viewing, and our webinar funnel guide puts these tests in order across the whole funnel.
Use the metrics guide to choose the business outcome, the tracking plan to connect it to the CRM, and our PostHog experiment guide to run the test.