EtonHouse International Education Group AEO Assessment Report
by Novastacks AI

This is the strongest technical position in the comparison set, and it is worth stating first because it removes the usual blocker. All seven sampled EtonHouse pages are server-rendered on WordPress and fully readable by crawlers without JavaScript execution — there is no client-side rendering problem to solve, so the score is not capped.

On speed, EtonHouse leads both rivals. The homepage records FCP 1.2 s, LCP 1.6 s, CLS 0.001 and 0 ms of total blocking time, against Stamford American's 1.7 s / 1.9 s and the Australian International School's 1.4 s / 2.0 s. Performance scores 88 to their 81 and 85; accessibility scores 100 to their 89 and 94; search-engine best practice scores 100.

Two hygiene defects sit inside this otherwise clean picture and both are live in the sitemap: a misspelled curriculum URL and a misspelled careers URL. The first one matters more than it looks — when ChatGPT answered the branded question about EtonHouse, one of the ten sources it pulled was that misspelled page, presented to the reader with a misspelled title.

The priority pages are the broken ones. The China Immersion hub carries three top-level headings and not a single second-level heading; the Beijing destination page carries seven top-level headings and no sub-structure at all. A machine reading those pages sees seven competing page titles rather than one topic with named sections for pathways, destinations, preparation, family voices and FAQ — which is exactly the structure the page already has visually.

Three further defects sit on high-intent pages. The How To Apply page's main heading reads 'Generic Template', its title tag ends in a dangling dash and its meta description is blank — this is the page a parent lands on when they are ready to enquire. The homepage title tag also ends in a dangling dash with no brand name after it. The Newshub index exposes exactly one second-level heading and it reads 'Posts navigation'.

Both rivals do the opposite on comparable pages: Stamford American's Early Years page carries one heading and thirteen sub-headings; the Australian International School's curriculum page carries one and eleven.

All seven sampled EtonHouse pages carry question-and-answer content. None of them carries the markup that tells a machine those are questions and answers. The gap is total and it is the clearest single explanation for why AI systems describe EtonHouse from prose it has to interpret while it quotes rivals from data they publish.

EtonHouse ships site-level markup well — breadcrumbs on all seven pages, site and search markup on all seven, business-listing markup on all seven. What is absent is every type that identifies the organisation as a school and prices its offer: both rivals carry school and organisation identity markup on 7 of 7 pages and fee-offer markup on 7 of 7, while EtonHouse carries neither on any page. Stamford American and the Australian International School each also ship question-and-answer markup on two pages.

The business consequence is narrow and concrete: a 3,900-word answer library already exists on your domain, and machines are not being told it is an answer library.

We ran three core category queries plus the six parent-intent prompt pages as a signal test — nine live result pages in the Singapore market. Treat them as diagnostic signals rather than a census of the category.

The signal is consistent. On the category head term, the page-one result set is a mix of rival schools, directories and parent listicles — Stamford American at #3, the Australian International School at #6 — and EtonHouse does not appear anywhere in the top 23. On the broader preschool term, EtonHouse enters at #12, and it enters with the homepage rather than with a preschool page, which means the ranking page cannot answer the question that produced the search.

The inverse holds where EtonHouse has built something specific. On the China immersion query, the hub page ranks #2 and Google generates an AI Overview above it — the only one of the three category queries where an AI Overview fired at all. The category head term does not trigger one; the specific programme term does. That is the shape of the opportunity: EtonHouse is losing the generic ground and holding the specific ground, and the specific ground is where the AI answers are being generated.

Read the top 150 ranked keywords by traffic and the picture of who this site reaches is unambiguous. The single highest-traffic term is 'educational assistant' — 8,100 monthly searches, position 6 — and it lands on a job posting. The next tier is generic parenting: 'parenting' at #4 on 2,900 searches, 'bullying' at #7 on 1,900, a cluster of Chinese New Year explainers. Enrolment intent appears further down: 'pre nursery singapore' at #8, 'igcse schools singapore' at #10.

Two things follow. First, the parenting and editorial content is doing real work — it is the second-largest traffic source on the domain — yet the main parenting library sits on a separate subdomain where none of that authority reaches the pages that take enquiries. Second, and more directly relevant to the China Immersion brief: not one China-immersion or Mandarin-immersion term appears in the top 150 keywords by traffic. The highest-placed Mandarin term is 'mandarin preschool' at position 5 on 260 monthly searches, and it lands on a campus page. The hub ranks #2 on its exact-match query and contributes no measurable traffic, because nothing above that exact phrase — the suitability, safety, destination and cost questions that precede it — is being captured.

The two rivals are not the same problem. Across the 67 keywords EtonHouse shares with the Australian International School, EtonHouse ranks higher on 40 and lower on 27 — EtonHouse is ahead on that overlap. Across the 56 keywords shared with Stamford American, EtonHouse ranks higher on 15 and lower on 41.

The losses concentrate in one place: the international-preschool and IB vocabulary. Stamford American holds #1 on 'international preschool in singapore' to EtonHouse's #7, #4 on 'international preschool singapore' to EtonHouse's #28, and #12 on the IB primary-years term to EtonHouse's #38. That is the vocabulary a parent uses before they know which school they want — and it is the vocabulary that feeds the comparison prompts where EtonHouse fell outside the top 20.

Stamford American also converts more from less: 607 ranked keywords producing an estimated 23,419 monthly visits, against EtonHouse's 506 keywords producing 6,864. The difference is not keyword count, it is which keywords and which page answers them.

EtonHouse holds 37 keywords at position 1 and 34 more at positions 2-3 — more first positions than either rival (Stamford American 32, the Australian International School 30). Those 34 second and third positions are the most leverage in this report: each is one place from the top of the page, on a domain that already ranks.

Three clusters are genuine assets. The IGCSE and A-Levels cluster: EtonHouse holds #10 and #11 on two 3,600-search terms where the Australian International School sits at #19 and #23. The early-care cluster: #12 and #13 on 'infant care' and 'infant care singapore' against #20 and #29. And the age-eligibility cluster — 'nursery age singapore', 'singapore preschool age' — where EtonHouse holds positions 7 and 8 against #39 and #20. These are decision questions with commercial intent behind them, and EtonHouse already owns the answer.

The China Immersion position belongs here too: #2 organically on its category query, with Google generating an AI Overview on that query and ChatGPT leading its answer with EtonHouse's own comparison of Beijing, Chengdu and Suzhou. The programme is already the most AI-legible thing EtonHouse publishes. It is also the thinnest-linked.