Discussion starter: a content SEO question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.
- AI Search
- Content SEO
Universities lose ‘study X’ queries to course aggregators and AI answers. Has any institution won them back, and what changed on its course pages?
Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.
1 reply
-
adminAdmin
Public evidence is thin. Well-documented cases of a university taking generic “study X” head terms back from course aggregators are hard to find, and a recent agency case study with real detail (below) reports gains in AI citations for course-level queries rather than a return to the top of the bare subject query. That does not make the work pointless. It reflects intent: a single course page is a structurally weaker answer to “where should I study nursing?” than a page comparing every provider, and better copy does not change that.
The more useful question is which parts of “study X” demand an institution can win. It has the stronger answer when the query carries something only it can satisfy: a location, a mode such as online or part-time, an entry route, an accreditation, or the facts about its own courses that aggregators and AI answers repeat second-hand. The course-page changes that matter are the ones that make those facts complete, current and quotable, on one stable URL.
What public evidence exists, and how to read it
A recent example with specifics is IFACTORY’s September 2026 case study of an anonymised US regional public university with about 5,000 students. The changes it describes:
- Title tags that lead with the program name.
- An opening paragraph of 150 to 200 words stating the credential, the program, the institution and the location.
- Scannable sections for admission requirements, program length, cost, careers and how to apply.
- The institution named in the H1 on tuition and admissions pages.
- FAQ schema answering student questions.
- Legacy PDFs that outranked recruitment pages removed or noindexed.
Reported results, comparing 2025 with 2026: AI mentions rose from 502 to 724, AI citations from 538 to 833 and cited pages from 401 to 455, and traffic from AI search grew 769% year on year in July.
Treat it as a single agency-published case study, which sits low on the evidence ladder:
- No control group. The write-up itself notes it covers one institution over one academic cycle, during which AI search itself changed. AI referral traffic grew across the web over the same period (BrightEdge, for example, reported ChatGPT referral traffic more than doubling in 2026 without detailing its method), so a large percentage from a small base says little on its own.
- A different outcome. The headline figures are AI visibility and traffic measures, not rankings for generic subject queries or application numbers.
- A dated tactic. FAQ rich results stopped appearing in Google in May 2026, so any benefit from the FAQ work now has to come from the visible questions and answers, not the markup.
What transfers is the pattern. Every change either made a fact explicit and attributable (credential, institution, place, cost, entry requirements) or removed an older document competing for the same queries.
There is also a reason to think aggregator positions are less fixed than they look. Aleyda Solis’s analysis of the May 2026 core update, based on Sistrix visibility data for the US and UK, concluded: “For affiliate, comparison, reference and tool sites: a derivative layer now needs a much stronger reason to exist in the SERP.” It also found that category-defining marketplaces, where users can complete the task, gained. The analysis did not examine education, so treat it as a hypothesis to check in your own subject results, not evidence that course aggregators lost ground.
Why aggregators have the edge on the bare query
The bare query names no institution, so the searcher is at the comparison stage. Think of a course page as a shopfront and “study nursing” as a question asked at the shopping centre’s directory board: the directory wins that moment because it lists every shop, even though each shop knows its own stock better. In search terms, ranking systems try to match the intent behind a query, and for a comparison-stage query a page that lists many providers, with filters for location, entry score, mode and cost, fits that intent better than any single provider’s page, however good.
AI answers take the same query in another direction. They summarise what the degree involves, typical entry requirements and careers, and may name institutions. Those facts come from somewhere, and for its own entry requirements, fees, structure and intakes, the institution is the primary source. Much of what aggregators show about a course is supplied by, or restated from, the institution. When the institution’s own page states those facts poorly (fees in a PDF, entry requirements on a separate admissions site, a marketing opener with no facts), the second-hand version becomes the easier one to find and quote.
The parts of “study X” an institution can realistically win
Query set Illustrative query Why aggregators have the edge Where the institution has the edge Bare subject “study nursing” Breadth, comparison and filters across every provider Weak unless the institution leads the subject nationally; treat it as awareness Subject and place “study nursing in Perth” A list of every local provider Campus, local placements and local entry routes, stated on the page Subject and mode or level “part-time nursing degree”, “graduate entry nursing” Filters across providers A smaller pool of providers, and details only the course page holds Subject and entry route “nursing without an ATAR”, “mature age entry nursing” Generic explanations The institution defines its own pathways and can state them with authority Subject questions “how long does it take to become a nurse” Explainers, plus AI answers that may answer directly A page that states the answer for its jurisdiction and courses can be the cited source Institution and subject “nursing at [university]” None Should belong to the institution; if an aggregator outranks you here, fix this first What to change on course pages
Element What good looks like Failure pattern to check for URL One evergreen URL per course, with the entry year handled on the page A new URL each intake year, with last year’s page redirected, orphaned or returning a 404 Variants One page per course, with domestic and international fees and entry requirements as labelled sections unless the offers differ materially Near-identical pages per student type or study mode, splitting links and relevance Opening paragraph Credential, subject, institution, campus or online, duration and next intake in two or three plain sentences A marketing opener with no facts Entry requirements and fees In HTML, dated (“for 2027 entry”), with the basis of any entry score explained In PDFs or images, or on a separate admissions site the course page does not link to Structure Units or modules by year, with enough description to show what is distinctive A link to a handbook PDF Outcomes Professional accreditation named; graduate outcomes from an official survey (QILT in Australia, Discover Uni in the UK) with the survey year “Excellent career prospects” with no data People and proof Academics’ expertise, placement partners, real student work Stock photography and generic testimonials Competing documents Old prospectuses, handbook entries and PDFs checked against the queries the course page should own A years-old PDF or handbook page outranking the current course page Where a university runs a separate handbook or catalogue system alongside its marketing course pages, the two can compete for the same entry-requirement and unit queries. Decide which one is meant to rank for what, and link them to each other.
Do not expect markup to do the heavy lifting. Google phased out the Course Info rich result in its June 2025 simplification of search results, and its guide to optimising for generative AI features says special schema is not needed for them. Keep accurate markup if your platform outputs it, but the changes likely to matter are in what the page says, not how it is tagged.
Build a subject hub, not only course pages
The page that can compete for subject-level demand is a subject hub: one page per major subject covering every level and route the institution offers (undergraduate, honours, postgraduate, online, pathways), explaining which course suits which kind of student, and linking to each course page. It answers the comparison question within your own portfolio, which is the part of the comparison you can own honestly, and it gives the course pages a strong internal parent. It is also the natural home for the jurisdiction-specific facts a comparison-stage searcher needs, such as the route to professional registration in that field.
Hypothetical example:
A regional university segments its nursing queries across one recruitment cycle. The bare “study nursing” term shows aggregators and an AI Overview at the top, with the university’s course page on page two. “Study nursing in [city]” has the university in the top three. “Graduate entry nursing [state]” puts the course page mid page one, and questions such as “how long does a nursing degree take in [state]” are answered in AI Overviews that cite an aggregator. The decision: build a nursing hub aimed at the city and state queries, move the course-length and entry-route facts into HTML on the course pages so they can be cited, and report the national head term as an awareness metric rather than a target.
How to tell whether you are winning anything back
- Split the query sets before comparing anything. Branded, institution-plus-subject, subject-plus-place, subject-plus-mode and question queries behave differently. Search Console’s branded queries filter separates brand demand, and regex handles place and mode. Query filters such as “queries containing” drop anonymised queries from the totals, so regex-built sets will not add up to the property total.
- Compare like-for-like points in the recruitment cycle. Course demand moves with the admissions calendar (offer rounds, clearing, application deadlines), and those dates shift from year to year. Compare against the same stage of the previous cycle, and judge the work over at least one full cycle, because that is the period over which demand repeats.
- Track AI visibility by page. The “Search generative AI performance” report gives AI Overview and AI Mode impressions by page, with no clicks or queries, and it has been available to all eligible sites since 31 August 2026.
- Measure applications, not only rankings. Track enquiries and applications by course and source. A student who met you on an aggregator and then searched your name shows up as branded demand, so falling generic rankings alongside rising branded searches and steady applications is not a loss.
- Audit your aggregator listings. If aggregators rank, they are part of your funnel. Check that their facts match your course pages and that they link to the canonical course URL.
When to work through aggregators instead
Competing and listing are not exclusive. Where searchers compare countries or dozens of providers, international postgraduate searches being the obvious case, aggregators are part of the path, and the practical job is complete, accurate listings that point to canonical course pages. Where the institution’s offer is distinctive in place, mode, entry route or accreditation, compete directly. And where undergraduate applications run through a central service (UCAS in the UK, state admissions centres in Australia), the task-completing step sits with that service rather than with you or any aggregator, so the course page’s job is to win the choice before the application.
A practical position
If I were running this for a university, I would stop treating the bare “study X” term as the target and report it as awareness. I would fix course pages so every fact a comparison-stage searcher or an AI answer needs is in HTML, dated, attributable and on one stable URL; build a subject hub for each major subject; and retire the old documents competing with both. Then I would judge the work over one full recruitment cycle, by query set and by applications, against the same point in the previous cycle. If the place, mode and entry-route query sets grow and applications hold, the work is paying off, whatever the head term does.
Need help with this on your own site?
Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.
Add a reply Cancel reply
Replies are open to approved members. Log in or apply to join.