SEO For Universities: The Estate Level Work That Moves Rankings

SEO for universities is usually run as a keyword exercise on a marketing site, and it usually stalls. Not because the tactics are wrong, but because the ranking page is not the page the team controls, the fact that decides the query is not on it, and the same course exists at four other URLs across the estate.

A university estate is an unusual search problem. Thousands of pages, dozens of publishing teams, several platforms, a seasonal traffic cycle with a hard deadline, and a set of queries where a third party aggregator holds better structured facts than the institution does. Fixing that is technical and editorial work on templates, not a content calendar.

Our August 2026 audit of 64 UK higher education websites measured what those estates publish in machine readable form. Read as a search backlog, the findings say the same thing repeatedly: the decisive facts are not in the HTML that gets indexed.

Start from the query types, not a keyword list

Applicant demand splits into six recognizable groups, and each one belongs to a different page. Mixing them is how an estate ends up with a subject hub that ranks for nothing and a course page that answers nothing.

Table mapping six higher education query types to the page that should own each one and the usual failure

Two of these rows explain most of the lost demand. Cost and eligibility queries carry the highest intent in the sector, and they are the two fact sets institutions are most likely to keep off the page, behind a filter or inside a PDF. Comparison queries are conceded almost entirely, which is why a league table site is often the first result for an institution's own course.

Brand and reputation queries are the ones estates do rank for, and they are the least decisive. An institution can hold position one for its own name and still lose the applicant on the course page.

The audit findings are a search backlog

Every measurement below was taken from the live estate, not from a survey. Each one has a direct search consequence.

Six audit findings from 64 UK institutions with the ranking and citation consequence of each

The first three rows are the same defect at different heights: facts that exist for a browser and not for an index. The detail of how to correct them sits in publishing course facts so answer engines can read them and structured data for university course pages.

The crawl rows matter for a different reason. When 92 percent of institutions name no AI crawler in robots.txt while 11 percent block those crawlers at the network edge anyway, the estate's exposure to AI answers is accidental rather than chosen, which is the argument in AI crawler policy for higher education.

The biggest ranking problem is internal

Before looking at competitors, count how many URLs in your own estate can answer one course query. In most institutions the answer is four or five.

Diagram of five URLs in one university estate competing for a single course query, with the role of each

The prospectus PDF is the one nobody wants to touch. It is a marketing artifact with an owner, a print schedule, and a habit of ranking for course queries it cannot serve. A PDF cannot be kept current, cannot carry structured data usefully, and cannot lead to an application without a detour.

Campaign and clearing pages are the second recurring cost. They are built fast, indexed permanently, and rarely redirected when the campaign ends, so the estate accumulates seasonal duplicates that compete with the pages meant to own the query.

Resolution is unglamorous: choose the owner page per course, consolidate or redirect the rest, point every internal route at the owner, and stop indexing thin finder result pages. Thirteen percent of audited institutions run a split course estate, where this work is a platform conversation as much as a search one, as covered in composable DXP programs.

What technical SEO means on a university estate

Server rendered facts. If fee, entry requirements, duration and start date are not in the initial HTML, no other tactic recovers the query.
One canonical course URL per award and study mode, with parameters and finder results excluded from the index.
Course and organization structured data mirroring the visible text, not a richer version of it.
Internal links from subject hubs, related courses, and the finder to the owner page rather than to a search result URL.
Performance measured on a mid range phone. A course page that takes six seconds loses the session before it can convert.
Redirect hygiene through the seasonal cycle: clearing, open days, campaign pages retired properly rather than left indexed.
Crawl budget spent on course pages instead of thousands of filtered finder combinations.

Most of that list is delivered in templates and the content model, not in page edits, which is the point of university website design as estate design. And several items are the same work as accessibility conformance, covered in WCAG 2.2 for university websites.

Measure the estate, not the keyword

Rank tracking a keyword set tells you very little in a split estate. Three measures tell you more: the share of course pages carrying fee and entry requirements as text, the number of URLs able to answer each course query, and the proportion of course pages with structured data matching what is on the page.

Those three are countable, they move with template work, and they explain a ranking change rather than reporting one. Add a fourth if you can: how often an AI answer about your courses cites your own domain rather than an aggregator.

Fixing the estate in the order that compounds

Search work on a university estate fails when it is sequenced as a keyword programme. The pages are owned by faculties, the facts are owned by admissions, and the templates are owned by digital, so a recommendation list lands on three teams who each need someone else to move first. Sequence the work by dependency instead.

Establish one owner page per course, and decide what happens to the duplicates: consolidate, redirect, or de-index the finder result pages that compete with it.
Put the deciding facts into that page as server rendered text, from a source that updates itself.
Fix the template, not the page. Headings, internal links to funding and accommodation, and a single canonical pattern applied across the course template group.
Only then work on titles, descriptions and content expansion, where the effort is per page and the return is smallest.

The reason this order matters is that steps one and two also fix answer engine visibility, accessibility and campaign conversion, so the same effort is paid for three times. Starting at step four produces a report that shows movement on ranked keywords while enquiries stay flat, which is the outcome most institutions have already bought once.

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 estate problems in this article were the starting conditions there: disconnected touchpoints across admissions, research and student services, and manual workflows for course and funding updates. Content was rich and context poor, so pages competed with each other and none of them held a complete, current set of facts.

Two structural fixes did most of the search work. First, content as a service: modular content blocks reused across course pages, articles and campaign landing pages, so one owner page per course could carry consistent facts instead of four partial versions. Second, automation for course data, fees, funding and application deadlines, which removed the manual step that made pages go stale between cycles.

Currency is the part search teams underrate. A course page that is correct in November and wrong in February loses on both fronts: rankings drift to aggregators holding your figures, and answer engines quote the aggregator. Automated fields keep the owner page the most accurate copy of the fact. See the University of East London case study for the architecture.

Frequently asked questions

Where should a university start with SEO?

With the course template. Get fee, entry requirements, duration, award and start date into server rendered HTML on every course page, then resolve duplicate URLs for the same course. Both changes apply estate wide from one piece of work.

Do prospectus PDFs hurt search performance?

They compete with the pages meant to rank, and they cannot be kept current. Keep a downloadable prospectus if recruitment needs one, but exclude it from the index and make the HTML page the canonical source of every fact it contains.

How do we handle multiple domains for one institution?

Decide which domain owns the course journey and consolidate towards it. Where a second domain has to stay, remove the duplicate course content from it and link to the owner page rather than mirroring the facts.

Is SEO still relevant when applicants use AI assistants?

More so, because the same substrate feeds both. An answer engine quotes facts it can extract from your HTML. If the fee is only in a script, neither a search result nor an AI answer can carry it.

Who should own university SEO?

A named owner in digital or web services with a route into the platform team, because most of the fixes are template and content model changes. Marketing alone cannot deliver them, and neither can engineering alone.

Where to start

Take one high demand course. Search its exact name, list every URL of yours that appears, then check whether the fee and entry requirements are text in the source of the page that ranks. That five minute test usually produces the first quarter of the roadmap.

The audit view of the sector is in the UK higher education AI discoverability report, and our search and growth work runs this alongside platform engineering rather than as a reporting layer on top of it.

SEO For Universities: Fixing The Estate, Not The Keyword List