The website was never the problem

The University of Calgary’s digital team brought UX Guys in in 2015 to start a redesign, beginning with the new student experience; really, the front door to the institution.

The stated problems were the ones you’d expect. The site was visually and structurally dated. It wasn’t mobile-friendly, at exactly the moment mobile had become the default interface for the audience. And it spoke in internal jargon: the language of an institution talking to itself, not to a seventeen-year-old and their parents working out whether UCalgary was the right place to go.

All three were real. None of them was the problem.

Applying, being admitted, and enrolling crosses recruitment, admissions, the faculties, financial aid, and the registrar. Each of those units had its own content, its own authors, and its own reasons. The jargon wasn’t a copywriting failure. It’s what a site sounds like when its structure mirrors an org chart and nobody is accountable for the handoffs between the boxes.

That distinction set up everything after it. Fix the words and the site regresses in eighteen months, which is the fate of most university redesigns. Fix who owns the words and it might hold.

We ran the same iterative design loop several times — research, workshops, design, usability testing — for New Students, then Current Students, Alumni, Faculties and Departments, and Administrative Services. Four years. What mattered was the order we did it in and the thing we built underneath it.

We interviewed first-year students, parents and staff about applying, admissions, and enrolment, and built journey maps of the pain points, offline and on. The population we actually wanted was prospective students: high schoolers who could go anywhere. They’re hard to reach, and we were starting in summer. So every person we interviewed had already gotten in. Friction remembered by someone who pushed through reads as an annoyance; the same friction experienced by someone who quit is a wall. We made the trade knowingly, which is what kept us from over-trusting the maps.

New Students was a key segment but also one with the clearest external user and the least internal ownership; almost nobody inside the institution had a proprietary claim on how a prospective student experiences applying. That let us prove the method on neutral ground. Walking into the faculty conversations later with a working example is a completely different negotiation than walking in with a proposal.

Each group controlled and authored its own sub-domain, and that was never going to centralize, nor should it have. They know their subject matter. So the design system, built as the work progressed, became the actual mechanism: units keep control of what they say, the system constrains how it’s structured and expressed. Not a component library. A governance instrument that let nobody lose authorship and the user still get one coherent institution.

The hardest group was the Cumming School of Medicine, well-funded, its own digital team, its own agency, its own budget. Their objection was legitimate: their audience is researchers, clinicians and donors, not eighteen-year-olds. What landed wasn’t a workshop. We were upgrading the Drupal environment and the old one wouldn’t be supported after migration. Ongoing support meant adopting.

The audience-based structure is still the top-level navigation of ucalgary.ca: Future Students, Current Students, Alumni & Donors, Faculty & Staff, About. Content has turned over many times in eight years since full migration. The organizing structure hasn’t.

The design system is still in place across the faculties, Cumming included, with no drift. It lives publicly at web.ucalgary.ca, where it’s been maintained and updated by people I’ve never met.

Most redesigns of this size are visibly eroding inside two years. This one absorbed eight years of staff turnover, budget cycles, and new communications leads with new opinions.

One caveat worth stating, because it’s the honest reading: part of why it held is that the design system was welded to the platform everyone had to be on. That coupling was the design decision, not an accident.

Here’s the unglamorous part. The workshops didn’t win Medicine over. The migration deadline did.

In a federated organization, alignment work builds willingness and structural decisions build compliance, and you need both. The workshops weren’t wasted, they’re why the units that came along didn’t resent it, and a large part of why the thing survived. But willingness alone wouldn’t have gotten the faculty with its own budget. Every case study claims the workshop was the turning point. Mine wasn’t.

The other thing I carry forward: when you can’t win control of a federated service and shouldn’t want it, design the constraint layer instead. People will accept limits on expression far more readily than limits on authorship.

What I’d do differently is fight harder for the prospective-student research, or schedule the work so that population was reachable. We traded it for the calendar. The calendar was the more negotiable of the two.