CommonSpirit Health · 2024–2025
The question was never whether patients could find care. It was how many gave up first.
Twelve behavioral archetypes and one connected search system for doctors, locations and content across 2,200 care sites.
- My role
- Experience and design lead. Led the UX team: journey mapping, archetypes, prototyping, test design and client management on a consulting engagement.
- Team
- Consulting engagement; our team delivered during CommonSpirit's brand transition.
- Constraint
- Fragmented search patterns, inconsistent physician data, and earlier efforts that had stalled on the number of variables.
- Result
- +9.2% search completion (observed at six months)
Design one search system from the most common real journeys — usually a patient with a provider or place already in mind.
- +9.2%Search completionObserved at six months
- +6.3%TrafficObserved at six months
- $15.9MBusiness valueEstimated
Completion and traffic observed six months after relaunch. Business value is an estimate.
Evidence
Stakeholder interviews named the pain points; Adobe and Google Analytics, CRM conversions and EHR volume confirmed them. Analytics also showed appointment setting as the next step, and where most people dropped out.
A competitive audit showed where other health systems did better. We overlaid it with decision points and component requirements to get a shared roadmap with scenarios and edge cases to test against.
Strategic choice
Personas were too narrow for the range of motivations, so we defined 12 archetypes from real interactions. We tested interactive prototypes for searching, filtering and comparing, at scale and in moderated sessions.
What we found shaped priorities: most searches started from a physician or facility; without geolocation, search returned nothing and offered no fallback; slots shown as available were sometimes revoked; too many filters overwhelmed people and too few left them stuck; ratings mattered.
Prioritization
Filters adapted to the path a person took and included details like spoken languages. If a doctor wasn't taking new patients, we surfaced others at the same facility and nearby practices meeting the same criteria. A dead end became a next step.
We extended the same components to locations, the conditions library and utility pages through a shared taxonomy, so search behaved the same everywhere.
Execution
CommonSpirit was midway through a new brand, so we built Brand Center first — a smaller site that gave the brand a home and started the design system before the flagship relaunch.
Testing also showed mobile search was often driven by place and time, and that people across age groups welcomed AI-driven scheduling if it was fast and forgiving.
Limits
- The backlog includes more accurate physician data, a third-party ratings API, broader use of EHR data and deeper conditions content.
- Some doctor and location page sections remained blank at relaunch.