Most Schema Markup Is a Waste of Time. These Types Actually Work

Plenty of the schema people add in 2026 buys them nothing: types Google quietly stopped showing as rich results, markup duplicated by two plugins at once, and fields filled in for machines that never look at them. A smaller number of types still earn real placement and still tell AI answer engines exactly what your page is. This is which is which.

Schema markup is structured data — usually JSON-LD embedded in your page's HTML — that tells machines explicitly what a page contains: an article by this author on this date, a product at this price, a business at this address with these hours. In 2026 it earns you two things: eligibility for rich results in traditional search, and unambiguous machine-readability for the AI assistants that increasingly answer questions your customers used to type into Google. This guide covers which schema types are worth adding to a WordPress site, which have honestly lost SERP value, what the markup looks like, and how to implement and validate it without writing code.

Why structured data matters more, not less

A reasonable person might assume schema matters less now that AI can "just read" pages. The opposite is closer to the truth, for two separate reasons.

Rich results still work. Structured data remains how you become eligible for enhanced listings in Google — review stars, product details, event listings, recipe cards, breadcrumb trails. Eligibility is the honest word: Google decides what to display, and marked-up pages are candidates, not guarantees. But you cannot win a lottery you have no ticket for.

Machines reward the removal of ambiguity. AI assistants extracting answers from the web must decide what a page is, who wrote it, when, and whether its facts are structured enough to lift safely. Prose forces inference; JSON-LD states it outright. No AI provider publishes exactly how it weighs schema, and honesty requires saying so — but everything we know about retrieval systems says explicit beats inferred, and schema is the web's standard way of being explicit. It is a core pillar of answer engine optimization, alongside content structure and llms.txt.

One important limit up front: schema describes your content, it does not improve it. Marking up a thin page produces well-described thin content. Schema is a multiplier on substance, never a substitute.

The types worth adding — and their honest 2026 status

Not all schema is equal, and the SERP landscape has genuinely changed. Here is the frank picture, type by type. The full menu is in the schema types reference.

Article — add it everywhere, expect no visual reward

Article schema (with author, publish and modified dates, and publisher) rarely produces a visible SERP feature anymore. Add it anyway, to every post: it is the baseline machine-readable statement of authorship and freshness — exactly the signals both search engines and AI assistants use to decide whether to trust and cite you.

FAQ: no rich result left, structure still worth having

Honesty checkpoint: Google cut FAQ rich results back to a narrow set of government and health sites in 2023, then retired them for every site on May 7, 2026. If you were adding FAQ schema for SERP real estate, there is none left to win. Add it anyway where you genuinely answer questions on the page, because explicit question-answer structure is precisely the format answer engines extract from, and Google has said markup it no longer uses does no harm. Setup: FAQ and HowTo schema.

HowTo: rich result gone since 2023, structure still useful

Same story, three years earlier: Google removed HowTo rich results from mobile and desktop results in 2023. The structured step-by-step data remains a clean, machine-readable representation of instructions, useful for AI extraction but no longer a SERP feature to chase.

Local Business — high value, chronically under-implemented

If you serve a physical area, this is among the highest-ROI markup you can add: name, address, phone, hours, and geo details in canonical machine-readable form, feeding both local search and every assistant asked "is there a good ceramics studio near me that does classes?" Guide: local business schema.

Product — essential for anything sold

Price, availability and review data power some of the most valuable rich results still fully alive in the SERP, and product facts are exactly what shopping-oriented AI queries retrieve. If you sell, this is non-negotiable.

Breadcrumbs — small, cheap, keep it

Breadcrumb schema helps engines display and understand your site hierarchy. Minor, nearly free, no reason to skip.

Recipe and Event: strong where they apply

Both retain meaningful rich-result treatments and map to high-intent queries: recipe cards and carousels for Recipe, and event listings for Event, as long as the event happens at a physical venue (Google leaves out online-only events). If your content fits one of these, mark it up properly and completely rather than partially.

Course used to sit in this group. Google phased out the Course info rich result in June 2025, along with ClaimReview, Estimated Salary and a few others, so Course markup now describes a course for machines without earning a SERP feature.

What schema markup actually looks like

JSON-LD is the recommended and dominant format: a script block of structured JSON, invisible to visitors, sitting in your HTML. A realistic Article example:

json

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Schema Markup in 2026: The Practical Guide for WordPress",
  "author": {
    "@type": "Person",
    "name": "Jane Okafor"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Example Publishing"
  },
  "datePublished": "2026-01-14",
  "dateModified": "2026-08-30",
  "mainEntityOfPage": "https://example.com/schema-markup-guide"
}

That is the entire trick: facts you already display in prose, restated in a vocabulary machines share. The complexity in practice is not the format — it is generating correct, complete JSON-LD for every page automatically, keeping it synchronized as content changes, and never shipping malformed output.

Implementing schema on WordPress without code

You have three realistic options: hand-written JSON-LD in your theme (fragile, unmaintainable at any scale), a generic snippet plugin (better, still manual per page), or an SEO plugin that generates schema from structured fields. The third is the sane default, and it is how Hydrogen SEO does it:

Pick a type. In the Schema screen (site-level) or the post's SEO panel, choose Article, Product, FAQ, HowTo, Recipe, Course, Event, Local Business, Breadcrumbs and more. Getting-started walkthrough: schema markup overview.
Fill the fields. Each type presents its own fields. They support smart tags — type @ to insert dynamic values like the post title, author or dates — so one configuration stays correct across many pages.
Valid JSON-LD is output automatically, server-rendered into the page with zero front-end JavaScript, so every crawler and assistant sees it without executing anything.

Two accelerators worth knowing. For business schema specifically, AI-assisted schema generation can draft the markup from your details. And for bulk work, Hydrogen AI — the plugin's agent — takes plain-language jobs like "add FAQ schema to this page," reading the page's actual questions and answers and structuring them, asking before it writes anything. That is the agentic SEO model applied to structured data: you review markup instead of hand-building it.

Developers who need to customize output programmatically have real hooks — hydrogen_seo_schema_markup, hydrogen_seo_schemas and friends — documented in the filters reference.

The validation workflow

Invalid schema is worse than none: it can disqualify you from rich results and signals sloppiness to any machine reading it. Make validation a habit, not an afterthought:

Validate at publish time. Run new or changed pages through Google's Rich Results Test (for rich-result eligibility) and the Schema.org validator (for vocabulary correctness). Step-by-step: validate rich results.
Fix errors before warnings. Errors block eligibility; warnings are usually missing recommended fields — improve them opportunistically.
Check Search Console periodically. Its enhancement reports surface structured-data issues across the whole site, catching regressions your spot checks miss.
Audit site-wide. Schema health is one of the seven areas in Hydrogen SEO's site health score, and regular site audits keep drift from accumulating silently.
Never mark up what is not on the page. Schema describing invisible content violates Google's guidelines and can earn manual actions. The markup must mirror the visible page — always.

One more classic failure mode: two SEO plugins both printing schema produces duplicate, conflicting markup. Run one primary plugin, and diagnose duplicate output promptly if you suspect it.

A sensible rollout order

For a typical WordPress site, sequence the work like this:

Article schema on all posts, Breadcrumbs site-wide — the universal baseline.
Your business type: Local Business or Product, filled completely.
FAQ and HowTo where the content genuinely fits, purely for extraction value, since neither earns a rich result now.
Specialty types as applicable: Recipe and in-person Event for their rich results, Course as description only.
Validate everything, then fold schema checks into your monthly maintenance routine as part of the wider WordPress SEO checklist.

That is an afternoon of setup for most sites, and it permanently upgrades how every machine on the internet reads you.

Start with the ticket, then earn the lottery

Schema will not rescue weak content, and nobody can promise you rich results or AI citations. What it does is cheaper and surer: it removes every excuse a machine has to misunderstand you. In a world where both search engines and answer engines decide in milliseconds whether your page is safe to feature, that is a bet with no real downside.

Hydrogen SEO's schema tools are part of the free plugin — no account needed for any of it — with AI generation and the agent available on a free account's starting credits. If you are implementing schema across a larger site or several client sites, the pricing page covers heavier AI use. Pick your first page, add Article schema, validate it, and you have started.