We need to talk about how we handle data in WordPress. For years, the standard advice has been to just “spin up a Custom Post Type and add some ACF fields.” But if you look at how industries like healthcare handle information, you realize we’re essentially building digital junk drawers. They use Knowledge Graphs to create shared meaning, and it’s about time we stop treating our databases like flat spreadsheets.
I’ve spent 14 years refactoring “legacy” sites where the previous dev decided to store product dimensions as strings in one place and integers in another. This isn’t just a “clean code” problem; it’s a structural failure. Healthcare didn’t become a leader in data by having better servers—they did it by investing in shared ontologies and semantic infrastructure. They agreed on what things are before they started coding how they look.
Ontologies vs. The “Post Meta” Nightmare
In the WordPress ecosystem, we tend to think in terms of CPTs. In a Knowledge Graphs mindset, you think in terms of Entities and Relationships. Healthcare uses ontologies like Uberon or ChEBI to define every gene and chemical compound. They don’t guess; they use a controlled vocabulary.
In contrast, most WP sites are a mess of disconnected wp_postmeta rows. When you search for a “Product,” the database has no semantic understanding that the “Manufacturer” is a separate legal entity with its own attributes. It’s just a string in a column. This creates a massive bottleneck when you try to scale or integrate AI features.
The Naive Approach (The Mess)
Usually, a dev will just dump data into a meta field like this. It’s quick, it’s dirty, and it’s impossible for a machine (or a future developer) to reason about without knowing the exact “magic string” keys.
// The "I just need it to work" approach
update_post_meta( $post_id, 'med_info', 'Paracetamol 500mg' );
update_post_meta( $post_id, 'usage', 'Twice daily' );The Semantic Approach (The Solution)
If we want our sites to be future-proof, we need to anchor our data in open standards like Schema.org. By using JSON-LD, we turn our WordPress posts into nodes within a global web of data. Specifically, we can use a hook to inject structured data that search engines and AI agents actually understand.
<?php
/**
* Inject semantic JSON-LD into the head based on CPT data.
*/
function bbioon_inject_knowledge_graph_schema() {
if ( is_singular( 'treatment' ) ) {
global $post;
$schema = [
'@context' => 'https://schema.org',
'@type' => 'MedicalScholarlyArticle',
'name' => get_the_title( $post ),
'author' => [
'@type' => 'Organization',
'name' => 'Clinic Authority'
],
// This is where we link to a shared ontology identifier
'sameAs' => 'https://www.wikidata.org/wiki/Q37060'
];
echo '<script type="application/ld+json">' . wp_json_encode( $schema ) . '</script>';
}
}
add_action( 'wp_head', 'bbioon_inject_knowledge_graph_schema' );Lessons from Healthcare’s Semantic Playbook
Healthcare leads because they separated pre-competitive semantics from competitive advantage. Hospitals compete on patient care, not on the definition of a “patient.” We can learn a lot from this semantic infrastructure model.
- Standardize the Sharing: Don’t just standardize the data; standardize how it moves. Use REST APIs or GraphQL with strict schemas rather than proprietary endpoints.
- Treat Vocabularies as Infrastructure: If you’re building a directory, don’t invent your own category names. Use a shared thesaurus or external database identifiers (like GeoNames or LEI).
- Build Incrementally: You don’t build Knowledge Graphs overnight. Healthcare has been cataloging findings for centuries. Start by fixing one post type at a time.
Furthermore, aligning with these standards ensures your knowledge outlives any single theme or plugin. If you’ve ever had to migrate a site with 50,000 rows of unstructured meta data, you know the pain of not having a semantic anchor. Consequently, spending time on data architecture now prevents a “Race Condition” against technical debt later.
For more on how to clean up your data, check out my guide on banishing messy text.
Look, if this Knowledge Graphs stuff is eating up your dev hours, let me handle it. I’ve been wrestling with WordPress since the 4.x days.
Stop Digitizing Documents; Encode Systems
The healthcare industry didn’t succeed by just “digitizing paper.” They succeeded by encoding how the medical system itself works. When you build a WordPress site, you aren’t just making pages; you’re building a computable model of a business. If that model is broken at the database level, no amount of AI or “shiny” UI will save it. Refactor your data into a graph, stick to open standards, and stop building junk drawers.
