44.7% Form Start Increase On Acıbadem Patient Registration

Acıbadem worked with Webtures to review the patient registration funnel across high-traffic treatment pages end to end. The starting point was a gap between rising media spend and patient registration volume that was not growing at the same rate. The programme ran in two phases: behavioural research first, then a hypothesis-driven A/B testing programme.

Highlighted KPI cards

• Patient registration form start rate up 44.7%
• Form completion conversion rate up 8.2%
• Visitors seeing the form rose from 15% to 71%
• Form abandonment fell from 63.4% to 41.8%
• Median mobile completion time cut from 2:47 to 1:12
• Effective cost per patient lead down 30.8%

The conversion problem

In healthcare, conversion is not only a traffic or design question. A user landing on a treatment page has to clear two separate thresholds: reaching the action they came for, and trusting the brand enough to complete it.

These were defined as two distinct questions at the outset. Can users reach the registration form? Can those who reach it complete it? That separation made it possible to tell whether the problem sat in the design or in the placement.

Research approach

A mixed-methods approach was used. Quantitative data answered what was happening, qualitative data answered why.

• Funnel analysis: Page view, form view, form start and form submission were instrumented as separate measurable steps.
• Scroll and click maps: How far users scrolled and where they clicked without result was mapped.
• Field-level analytics: Abandonment was measured field by field.
• Session recordings: Recordings were independently coded and thematically analysed.
• Moderated usability testing: Task-based scenarios were run with participants matching the target patient profile.

Key findings

The form sat where most visitors never looked. The registration form was placed at the very bottom of the page, and scroll data showed only 15% of visitors reached it, meaning 85% of prospective patients never knew it existed.

The real insight was that 85% of users who did reach that section completed the form. Demand existed; the form was not where the demand formed. That was the strongest evidence that placement, not form design, was the bottleneck. A redesign started without scroll data would have spent the budget in the wrong place.

Length was not the issue; specific fields were. The form had 11 fields, and field-level measurement showed 72% of abandonment concentrated in just four: national ID, address, manually entered date of birth and a mandatory complaint description. Sensitive fields such as national ID and address produced a clear privacy reaction in usability testing, with 9 of 12 participants questioning why that information was required at that stage.

Interaction delay on mobile caused mistaken taps. 78% of traffic was mobile. Validation scripts firing on field focus pushed INP to 480 ms, and a layout shift of 0.31 on keyboard open made users tap the wrong field.

Trust signals were disconnected from the decision. Accreditation, medical staff and patient reviews were present on the page but not in the same viewport as the form, so users did not see them while deciding. In a five-second test, only 3 of 12 participants could clearly state that the page was a hospital appointment page.

Testing programme

Findings were turned into a prioritised hypothesis list and tested in sequence. Tests ran on server-side infrastructure with a 50/50 traffic split. The decision criteria were 95% confidence and at least 80% statistical power, with the required sample size calculated in advance and no test stopped before reaching it.

• Form placement: The form was moved to the point of decision while the existing footer form was kept. Form starts rose 38.9%.
• Field reduction and two-step structure: Step one was reduced to name, phone and department. National ID and address moved to the call centre process. Abandonment fell from 63.4% to 41.8% and completion rose 5.1%.
• Trust layer: A privacy notice, JCI badge and a 24-hour response commitment were added to the form card. Completion rose 2.4%.
• Mobile performance: Validation scripts were deferred, keyboard behaviour was corrected and a persistent appointment action was added. INP fell from 480 ms to 165 ms, CLS from 0.31 to 0.04 and LCP from 3.8 s to 2.1 s; mobile form starts rose 9.3%.

Outcome

By the end of the programme, form start rate rose from 2.9% to 4.2% (up 44.7%) and form completion conversion rate rose from 36.6% to 39.6% (up 8.2%). Effective cost per patient lead fell 30.8%. The additional volume did not reduce lead quality on the call centre side; the share converting to appointments held, meaning the increase reflected real patient demand.

The most transferable lesson concerns diagnosis. Identifying the bottleneck correctly mattered more than the fix itself. In healthcare, friction is often a trust problem, confirmed here by deferring sensitive fields to a later stage. Performance behaved not as a technical topic but as a user experience metric in its own right.