Twenty-two sites into one

Situation

Each of the 22 locations had its own marketing department, responsible for its own site’s technology, risk, experience and content. The brand looked different from site to site, and any work across them meant coordinating 22 teams.

The locations offer different specialties and different research into new treatments. With 22 separate sites, a family had almost no way to see what the system as a whole could do, and Shriners had almost no way to bring a patient from one region to a location in another.

Publishing was its own problem. Each team was stuck with a rigid process for writing, approving and publishing pages, the work competed with everything else on their plates, and a lot of the content was out of date.

What I did

I led discovery with the marketing VP at each location, physicians, the head of medicine and patients’ families.

The core of the design was calling things the same thing everywhere. Locations, specialties, conditions, research and treatments were all tagged one way, so a page about burn care could pull in the locations that treat burns, the doctors who do, patient stories and treatment options without anyone curating it by hand.

With that much content, families don’t always know which bucket to start in, so the site leads with search. An “I Want to Find” menu stays on every page, and the homepage opens with a search box that suggests where to go. Some actions also tag a visitor to an audience group, which changes the calls to action they see from then on.

For the marketing teams, the pages are built from components and a drag-and-drop page builder. Each component uses the same tagging, so it can be dropped on any page and fill itself with related content.

Governance was built into the CMS itself. I set up the roles, permissions and first assignments for each location’s team.

Wireframes came from a component library I built at Valtech, which reused production code for features like find-a-doctor and search.

What changed

One site replaced 22. Before-and-after analytics showed success rates for visitors who came with a specific task more than doubled. By the project retro, teams had cut the time to write, test, approve and publish a page by over 80%. The governance built into the CMS has held.

The search component went on to other clients. Tampa General Hospital used the same component, and that project reached market 40% faster. It was the first time we measured it, and we saw similar gains on a couple of later projects where other UX designers did the work.

What’s still hard

Twenty-two teams gave up control of their own sites for shared templates and permissions they didn’t write. That trade holds only as long as someone keeps tending the governance.