From one page to five: the content architecture missing from a school invisible in search
The brief was an on-page SEO audit. The diagnosis found something more serious: the entire site was a single indexable page, and "About", "Teaching" and "Contact" were anchors on the same URL. Without a content architecture, optimising titles and meta tags would be polishing a door that opens onto nothing.
Project delivered in partnership with Guiaponto, who brought the client and led the relationship. The technical front — on-page SEO and content architecture — is ours; credit for the account is shared.
The end client is a private school in Sorocaba, in the interior of São Paulo state, that did not appear in the first results for the search that matters most to it: parents looking for a private school in the city.
The diagnosis changed the size of the problem
The engagement was an on-page audit: title, meta description, headings, structured data. All of that was indeed missing — verified in the raw HTML, not estimated by a tool: zero meta description, zero structured data on the page, and no SEO plugin installed in WordPress, which explained all three at once.
But the finding that mattered was different. The sitemap confirmed it: the site had a single indexable page. The menu items "About", "Teaching", "School" and "Contact" were not pages — they were anchors scrolling within the same URL. That is not thin content, it is the absence of architecture. A search engine has nowhere to file each topic, and the whole site competes for everything from one address.
The real business objection was not technical
The company registry shows the school opened in August 2025 — less than a year of existence. Of the nine direct competitors mapped in the city, none has less than twenty years of history, and the oldest is past seventy.
No tag adjustment solves that. The content strategy had to answer head-on the question a parent asks silently — "why trust a new school?" — using the real signals that already existed. That is where the most strategic FAQ question came from, rather than from a generic list of frequently asked questions.
What was implemented, across five fronts
First: a 60-character title, a 146-character meta description and educational-organisation structured data with the real name, address and phone — delivered by a plugin written for the project, without installing any new SEO plugin on a site already running a cache layer and a visual builder.
Second: eight new content sections and the FAQ going from five to twelve questions, with FAQPage structured data.
Third, the one that changed the site: from one page to five. Home, who we are, what we teach, what we believe and frequently asked questions, at 1,022, 1,542, 2,221, 1,952 and 814 words. Sections were moved, never copied — duplicating content between the home page and the inner pages would have been worse than the original problem. The home page dropped from roughly 15,000 pixels of scroll to 4,743.
Fourth and fifth, already outside the contract: visual fixes across eleven buttons and thirty-eight text blocks, and accessibility taken into the green on PageSpeed, at the client’s request.
A regression I caused myself
While splitting the site, I wrote summaries that were too short for the home page hub. The home fell to 875 words and the highest-volume keyword — the one with 1,600 monthly searches — disappeared from it. Since the home is what competes for that search, this was a step backwards at exactly the point that matters most.
Fixed by rewriting the summaries, with the home back at 1,022 words and both keywords present. I brought no section back: duplicating would have solved the symptom and created another problem.
Optimising the title of a site with no architecture is polishing the sign on a shop that is all window and no aisles.
Two QA findings that stand on their own
The installed SEO plugin was marking the new pages as articles with a natural-person author — the wrong structured data for a school’s institutional pages. Default switched off, our own schema in its place, with breadcrumbs on the inner pages.
And the sitemap listed only the home page. The four freshly created pages had no way of being discovered by Google. Cache invalidated, all five URLs now listed. It is the kind of detail that shows up in no tool report and decides whether the entire job produces an effect or sits waiting for a crawl that never comes.
Deadline and scope
The contracted schedule was seven business days. The contracted front was completed in one. Everything that came after — content blueprint, multi-page architecture, visual fixes and accessibility — was additional scope delivered without additional billing.