Ten thousand pages, one front door.
A statewide academic medical center: eighteen hospitals, six colleges, and a research enterprise, living online as a dozen websites that had never met. We rebuilt the estate as one digital ecosystem and moved ten thousand pages through a single front door.
The challenge
On paper, the institution was one: a hospital system reaching eighteen hospitals across its state, a health-sciences university of six colleges, and a major research enterprise, all under a single name and a single strategy. Online, it was thirty-seven entities scattered across a dozen domains. Decades of independent site-building, compounded by rapid growth and acquisition, had produced college sites that never spoke to each other, health content arranged around the org chart, and no page anywhere that explained the whole. Every entity had a website. The institution didn't.
Discovery put numbers to the frustration. Fewer than four in ten employees felt the websites reflected the institution's mission, and only a third called them easy to use. Outside, the cost was starker: roughly half of prospective patients didn't consider the system an option for their family's care, and audiences consistently underestimated what it actually offered, from statewide clinics to research and clinical trials. Navigation and search topped every audience's list of complaints. The enterprise strategy said one institution; the web was the loudest thing contradicting it.
The approach
You can't migrate your way out of a structural problem. Before a single page moved, the work was to give the whole institution one mental model, an enterprise architecture where care, education, and research amplify each other instead of hiding from each other, and then to treat the migration itself as the enforcement of that model, one page at a time.
Baseline everything, then pick a north star
The discovery phase was deliberately exhaustive: fourteen sites audited across the dozen domains, thirty-three stakeholder interviews, surveys across seven audiences, and a round of workshops to pressure-test what we heard. It surfaced a clear diagnosis: a wealth of equally weighted content producing decision paralysis, and tasks that took circuitous paths from intent to action. It also surfaced a clear mandate, because stakeholders at every level were asking for the same thing. That became the strategy's spine: a unified, silo-free experience, organized around user need states rather than internal divisions, that works no matter where someone enters or who they are.
One enterprise IA, front doors instead of mazes
We designed a single information architecture spanning the entire enterprise: an institutional front door that establishes the whole, mission-area front doors for care, education, and research, and deliberate linkages between them, so that a patient reading about a condition can find the trial and a prospective student can see the hospital where they'd train. Underneath, the ecosystem shares its plumbing: unified provider, location, and program directories, a single content hub for stories across every entity, and a lean set of seven page templates replacing hundreds of one-off layouts. Landing pages became front doors that route people onward fast, not repositories that make them dig.
Move ten thousand pages, on purpose
The migration was where the architecture became real, and where it could have quietly died. I led the content workstream with a large content team behind me, and we ran it as editorial decisions rather than a bulk transfer: every legacy page audited, mapped to its place in the new architecture, and then kept, rewritten to the new voice and need-state model, or retired outright. Ten thousand pages earned their place and made the move, spanning health-system, college, and research content, on a governance model that made an owner and a job description explicit for every one of them. What launched was smaller, sharper, and entirely accounted for.
“I want a website that reflects who we are as one institution, not twenty different versions of us.”
What changed
The launch united all thirty-seven entities, from the eighteen hospitals and six colleges to the institutes and the enterprise itself, in a single digital ecosystem across a dozen domains. Seven distinct experiences shipped under one brand system: an enterprise front door, dedicated homes for health, education, and research, a shared system for the college sites, and flagship experiences for children's care and cancer care, each recognizably itself and unmistakably part of the whole. Directories for providers, locations, services, and programs finally worked across the entire system, and a shared content hub turned scattered news pages into one storytelling engine.
Just as important is what the migration left behind: a content estate where every page has a place, a template, and an owner. The old failure mode, where every entity built its own site because the whole had no room for them, now has a structural answer, and the team that maintains the ecosystem inherited a map instead of an archaeology project.
The outcome
The digital experience finally matched the enterprise ambition: one ecosystem, one brand, one front door, launched at full statewide scale.
- 0,000
- Pages migrated across healthcare and academic content for launch
- 0
- Entities, from hospitals and colleges to institutes, united in one digital ecosystem
- 0
- Distinct experiences shipped under a single unified brand system
What I’d carry forward
- 01
An enterprise IA is a political document as much as a structural one. The architecture only held because discovery built consensus first: every entity could see its place in the model before anyone was asked to give up a website.
- 02
Treat migration as ten thousand decisions, not one big move. The audit-map-decide pipeline is slower than a bulk transfer, but it's the only version where the new site launches lighter than the old one instead of inheriting its debt.
- 03
Shared plumbing is what makes 'one institution' true. Brand unifies the surface; unified directories, a common content hub, and a small template set are what make the ecosystem behave like one thing.
Have a system that has outgrown its structure?
This is the kind of problem I like best. Tell me where it stands and I’ll tell you how I’d approach it.