Building route-aware technical SEO for a multi-page corporate website
Metadata as architecture rather than a content field — generated per route, with structured data describing connected entities instead of isolated fragments.
Role
Architecture and delivery
Published
Reading time
1 min
JSON-LDStructured dataCanonical URLs
01
Context
A corporate website with a dozen or so pages, built to present an organisation to people who are searching for what it does. Standard requirements, standard scale, and the standard outcome — SEO treated as a content task performed once at the end.
02
Problem
Handled that way, the site ends up technically indistinct: several pages sharing a title pattern, no deliberate canonical strategy, social previews falling back to whatever text appears first, and structured data either absent or present as disconnected fragments that collectively describe nothing.
03
Constraints
01
Every route needs its own metadata
Not a template with a variable, but a real per-route title, description and canonical that cannot silently collide.
02
Structured data has to connect
A WebPage node and an Organization node that do not reference each other are two facts, not a description of an entity.
03
Two languages
Every decision has to hold in English and Arabic simultaneously.
04
Approach
I treated the metadata layer as part of the architecture. I generate titles, descriptions and canonicals per route from one source so divergence is impossible, and built the structured data as a connected graph — pages that reference the organisation, breadcrumbs that mirror the visible trail.
05
Architecture
Route definitionOne source per route for title, description and canonical.
Social metadataOpen Graph and Twitter tags generated from the same source.
Connected structured dataWebPage, BreadcrumbList and Organization referencing one another rather than standing alone.
06
Solution
Route-aware metadata generation, an explicit canonical strategy, complete social metadata, and JSON-LD describing the pages and the organisation as connected entities — implemented bilingually.
No ranking improvements, no traffic figures, no visibility claims. Those depend
on competition, timing and a search engine's own decisions, none of which a
build controls. What is verifiable is the implementation, and that is what this
page describes.
07
Lessons
01Generate metadata, do not author it per page. Anything typed by hand into a field eventually duplicates, and duplicates are invisible until something diagnoses them.
02Structured data is only useful when connected. Isolated nodes are decoration; a graph is a description.
03Breadcrumb markup must mirror the visible trail. Structured data that claims a path the page does not show is a defect, not an optimisation.
04Claims about ranking do not belong in a case study. What was implemented is verifiable; what it produced in a search index is not.