Rebuilding patient onboarding around one honest question
A 14-step signup flow became four screens, and activation doubled in a quarter.
- Engagement
- 16 weeks
- Team
- 1 strategist, 2 designers, 2 engineers
- Stack
- Next.js, TypeScript, PostgreSQL
- Year
- 2025
Results
- Increase in activation rate
- 2.1×
- Time to first appointment
- −68%
- Fewer support tickets in week one
- 41%
Meridian is a telehealth provider serving 400,000 patients across three countries. Their signup flow had grown one field at a time over six years, until new patients faced fourteen steps before they could book anything. Two thirds of them never finished.
The problem was sequencing, not length
Our first instinct was to cut fields. Watching eleven patients attempt signup told us something more useful: people did not abandon because the form was long. They abandoned because it asked for insurance details and medical history before showing them a single available appointment. The flow demanded commitment before it offered value.
The moment we saw someone quit at the insurance step and then say out loud 'I do not even know if you have anyone near me', the whole roadmap changed.
What we changed
We rebuilt around a single ordering principle: show availability first, ask for everything else later. Location and reason for visit came first, real appointment slots appeared second, and the clinical and billing detail moved to a post-booking step that patients complete while they are already committed.
- Four screens before a confirmed booking, down from fourteen
- Insurance verification deferred until after the slot is held
- Progressive medical history spread across the pre-visit week
- A resume-later link that restores state from any device
Building it without a rewrite
Meridian could not afford to freeze their platform, so we shipped the new flow as a parallel route behind a feature flag and moved traffic across in five percent increments. Every increment had a rollback path and a dashboard the clinical operations team could read without our help.
Twelve weeks after full rollout, activation had more than doubled and first-week support contacts fell by 41%. The clinical team now runs their own experiments on the flow using the same flag infrastructure we handed over.