University Website Design: What To Design Before The Homepage

Most university website design projects are scoped as a visual refresh and delivered as one. New palette, new photography, a rebuilt homepage, a navigation workshop. Two years later the same complaints return: applicants cannot find the fee, departments have gone their own way again, and the estate looks like six institutions rather than one.

The reason is that almost nothing an applicant needs is decided on the homepage. It is decided in the course template, in the fee table component, in the fields an editor is required to fill, and in whether a department can override the pattern. University website design is estate design, not page design, and the leverage sits below the surface most redesign briefs describe.

We audited 64 UK higher education websites in August 2026 to see what those estates actually publish in machine readable form. The findings are a design brief in disguise: 45 percent kept course fees out of server rendered HTML, 34 percent did the same with entry requirements, and 46 percent published no structured data on key pages. Those are template and content model decisions, made once and inherited everywhere.

Design for six readers, not for a brand audience

A university estate serves people arriving mid task from a search result, not visitors browsing a brand. Each group has a first question, and each leaves when the page answers a different one.

Table of six readers of a university website, the first task each arrives with, and what poor design costs each one

Two of those rows are usually missing from a redesign brief. Current students are the highest volume users of the estate and are routinely designed around rather than for, which is how a student ends up finding the submission deadline on a forum. And answer engines now read the estate on behalf of applicants, so a page that only renders its facts in the browser is a page that does not participate in the answer.

The practical move is to write the first task next to every template in the design system, then judge the template against that task rather than against a mood board. A course page that looks considered and does not carry the fee has failed its brief.

The course page is the whole project

If a redesign improves one template, make it the course page. It is where the decision is made, where organic and AI traffic lands, and where the estate's content model is exposed most plainly.

Side by side comparison of a brochure style course page and a decision oriented course page, with the design consequence of each element

The brochure column is not a straw man. It is the median UK course page: a large hero, institutional voice, rankings, modules hidden in a collapsed accordion, and the two facts that decide the application sitting behind a link to a central fees table or a PDF.

The decision column is not visually poorer. It is the same design language spending its first screen on four facts instead of one image. Once the facts are in HTML, the structured data can mirror them, which is the pattern covered in structured data for university course pages and publishing course facts so answer engines can read them.

Accordions deserve a specific ruling. A disclosure pattern is a legitimate design choice, but only when the content is present in the markup, keyboard reachable, and expandable without JavaScript rewriting the page. A module list that only exists after a click is a module list that is not published.

Where a design decision actually lands

The uncomfortable arithmetic of a large estate is that the same defect costs the same to design and a wildly different amount to fix, depending on the height at which you fix it.

Four layers where a design fix can land: design system, templates, editor workflow, and individual pages, with the reach of each

This is why a university design system is a governance instrument, not a component library. The contrast ratio, focus state, table pattern, and disclosure pattern defined once in the system are the same decisions that determine whether the estate passes an accessibility audit, which is the argument in WCAG 2.2 for university websites.

It is also where redesigns quietly fail. A system nobody is required to consume becomes optional, departments fork it, and the estate drifts back within two release cycles. The mandate matters more than the tooling, which is the operating model gap described in composable DXP programs.

The content model is a design decision

Design work in higher education stops at the wrong boundary. A course template can only render a fee if the content model has a fee field, and can only render it reliably if that field is required and typed rather than pasted into a rich text blob.

Fee, entry requirements, duration, award, start date and study mode as discrete required fields, not free text.
One source for each fact, referenced by every page that shows it, so a fee change is one edit rather than nine.
Editor guidance in the interface at the point of authoring, not in a style guide document nobody opens.
A validation rule that refuses to publish a course without the facts an applicant needs.

That last rule is the one institutions resist and the one that changes the estate. It converts a design intention into a published guarantee, and it is the reason a decision page stays a decision page after the launch team disbands.

What good looks like at estate scale

A design system with named owners, a version, and a consumption route for departmental sites.
Templates written against a first task, with the fact set they must carry documented alongside the visual spec.
One navigation for people mid task, tested with real queries rather than a card sort of the org chart.
Accessibility built into the component library, so a fix lands estate wide rather than page by page.
A performance budget per template. A course page that takes six seconds on a mid range phone loses the applicant before the design is seen.
A published scope for what the system covers, matching what the accessibility statement claims, as covered in accessibility statements in higher education.

What this looked like at the University of East London

The University of East London moved off Sitecore onto Drupal on Acquia with us, and rebuilt the estate as one platform rather than a set of pages. It is the closest reference point we have for this problem at full institutional scale, 125 years of content, students from 156 countries, and five distinct audiences reading the same site for different reasons.

The design work started with qualitative research and journey mapping across prospective students, current students, researchers, alumni and faculty, plus parents and counselors as a secondary audience. The mapping put a number on the pattern this article describes: users were dropping out between the research stage and the decision stage, on static, text heavy pages with no guided path from a course to funding to application readiness.

Three principles came out of that and governed every template:

Purposeful navigation: any course, research group or story reachable within two clicks.
Progressive disclosure: enough on the page to act, with depth available underneath rather than stacked on top.
Emotional visual language: inclusive imagery and clear typography carrying academic credibility without intimidation.

The measurable output we owned was sequence, not styling. Exits fell 46 percent and the institution reported a 73 percent rise in applications and a 32 percent rise in enrollments over the following two years. The applications figure is theirs to claim, ours was the journey structure underneath it. The full build is written up in the University of East London case study.

Frequently asked questions

How long does a university website redesign take?

A visual refresh of the main marketing site takes months. Re-designing the estate, meaning the design system, the course template, the content model and the editor workflow, is a program of a year or more, delivered in releases. The useful question is which template you improve first, not when everything ships.

Should the homepage lead the redesign?

Rarely. Most applicant sessions start on a course page from a search result or an AI answer, so the homepage is a small share of the decisive traffic. Improve the course template first and the homepage inherits the design language.

How do we stop departments breaking the design?

Make the compliant path the easiest one. A consumable component library, templates that already carry the required fields, and a review route beats a written policy. Where a department has its own platform, the system has to be shipped as code they can install rather than a document they can ignore.

Does design affect search and AI visibility?

Directly. Whether a fact is text in the initial HTML, whether a module list exists before a click, and whether a table is announceable are all design decisions, and all three determine whether the fact can be indexed and quoted.

Do we need a separate design system for student facing services?

No, but you need shared foundations and different templates. Marketing pages persuade and answer, service pages let someone complete a task quickly. Same tokens and components, different density and hierarchy.

Where to start

Open your highest volume course page on a phone. Count the seconds until the fee and the entry requirements are visible without a tap, then view the source and check whether both are in the HTML. That one test tells you whether your next design project is a refresh or an estate program.

If you want the outside view first, the UK higher education AI discoverability report shows what 64 institutions publish today, and our design practice covers how we run design systems and applicant journey work alongside platform engineering.

University Website Design: The Estate Decisions A Redesign Usually Skips