Schema Markup for Small Sites: Which Types Are Worth It?

This post may contain affiliate links. If you buy through one, we may earn a commission at no extra cost to you. See our Affiliate Disclosure for details.

Most one-person sites are already publishing structured data and their owners have never seen a line of it. An SEO plugin gets installed on day one, it starts wrapping every post in machine-readable labels, and that output sits in the page source for years without anyone opening it. Then a checklist somewhere says "add schema markup" and the whole thing gets a second layer bolted on top — duplicated, half-configured, occasionally describing a page that doesn't exist. The question worth asking about schema markup for small sites isn't how to add more. It's which of it does anything.

Markup is a label, not a ranking lever

Structured data describes a page in a vocabulary machines already parse. That's the whole mechanism. It doesn't argue that your post is good, and Google's introduction to structured data is careful to say that valid markup makes a page eligible for certain result types without promising it will get them.

Eligibility is a low bar with an unpredictable payoff. You can mark up a recipe perfectly and never see a recipe card, because the decision about what to render belongs to the search engine and it changes without notice.

So the honest framing is narrow. Markup buys you two things: a shot at a different-looking search result, and a page that machines read with less guesswork. Neither is a ranking boost, and treating it as one leads to sites where every element on the page has been labelled twice and nothing improved.

The types that still pay off on a small site

The full search gallery runs to dozens of entries — job postings, courses, vehicle listings, software apps. Nearly all of it is irrelevant to a blog or a service business. Four types cover almost every one-person site.

Breadcrumb

The clearest return on effort, because the result is visible: your search listing shows a readable path instead of a raw URL. Readers get context about where the page sits before they click. Practically every SEO plugin outputs this automatically once breadcrumbs are enabled on the theme.

Organization or Person

This is the sitewide "who publishes this" record — name, logo, social profiles, contact details. It doesn't produce a rich result by itself. It does give search engines a single consistent entity to attach your pages to, which matters more the smaller and newer the site is, since there's little else to identify you with.

LocalBusiness

Only relevant if people come to a physical location or you serve a defined area. Google's local business documentation covers hours, address, and service area. Keep in mind that a Business Profile is the heavier lever here — the markup supports what the profile establishes rather than replacing it.

Product and Review

For anyone selling something directly. Price and availability can surface in the result, which changes click behaviour more than most on-page tweaks do. One rule catches people out: reviews you write about your own business or products don't qualify. Self-serving review markup is disallowed, and marking it up anyway is a good way to lose rich result eligibility across the whole site.

Article markup belongs on blog posts too, but be clear-eyed about it. On a standard non-news blog it seldom produces a visible enhancement. It's descriptive plumbing — headline, author, dates — and it's already being emitted for you, which is the right amount of effort to spend on it.

FAQ and HowTo: the cautionary tale

For a few years, adding FAQ markup was the single highest-return trick in the SEO listicle economy. Pages doubled their vertical space in results. Then, in August 2023, Google narrowed FAQ rich results to a small set of well-known authoritative government and health sites, and cut back HowTo at the same time. Overnight, thousands of posts full of dutifully marked-up questions stopped rendering as anything.

The markup itself isn't harmful. Google's FAQPage documentation still describes the type, and if your plugin emits it there's no reason to tear it out. What's gone is the reason people added it.

Two lessons survive from that episode, and they're the reason this post keeps pointing at official docs instead of listing settings. Rich result types are withdrawn as casually as they're introduced, so any advice about them ages badly — including this. And a page structured around question headings because a rich result rewarded it, rather than because readers ask those questions, was always going to be left holding an empty format.

Structured data is the label on the port, not the signal running through it. Photo via Pexels.

Your plugin has probably done most of this already

Before adding anything, find out what's being published. Two free checks, both first-party:

Google's Rich Results Test takes a live URL and reports which rich result types that page currently qualifies for, plus the errors blocking the rest. It only reports on types Google supports, which is exactly the filter you want.

The Schema Markup Validator at schema.org is broader — it parses everything on the page against the schema.org vocabulary regardless of whether any search engine uses it. Use this one to spot duplicate blocks, which is the most common problem on sites that have been through more than one plugin.

Both of the major WordPress SEO plugins handle the standard types without configuration. Rank Math and Yoast SEO each build a connected graph of Article, Organization, Breadcrumb and WebPage data on every post, and both list schema output as part of their free tier on their plugin directory pages as of August 17, 2026. Feature splits between free and premium move around, so confirm the current line on the vendor's own page rather than from a comparison table.

What you should not do is hand-write JSON-LD into a theme file while a plugin is also generating it. That's how a page ends up declaring two different authors and three publication dates, and how the validator ends up flagging a site that was fine before anyone touched it.

Where markup goes wrong

Almost every real problem is a mismatch: the labels say something the visible page doesn't.

Structured data has to describe content the reader can actually see. Marking up a review that appears nowhere on the page, a price you don't charge, or an FAQ block hidden behind a toggle the visitor never opens falls under Google's spam policies, and the penalty isn't a warning email — it's silent ineligibility, often across the entire domain.

The second failure is subtler and much more common on small sites. Stale values. An author name from three redesigns ago, an address you moved out of, opening hours from before you went appointment-only. Nobody notices because nobody reads their own page source, and search engines keep repeating it back to the world.

A yearly look at what your pages actually claim is enough. Fold it into whatever pass you already do — an SEO site audit surfaces structured data errors alongside everything else, and the same afternoon you'd spend on an internal linking cleanup is plenty of time for both.

Whether structured data helps here is the open question of the moment — and nobody has published the rule. Photo via Pexels.

Does any of it reach AI answers?

This is where the marketing has run ahead of the evidence, so it's worth being blunt about what is and isn't known.

What's documented: structured data makes a page's key facts machine-readable, and language models ingesting web pages have an easier time with explicit labels than with inference from layout. That's a plausible mechanism, not a published rule.

What is not documented, by any major search or assistant provider: that adding schema markup causes a page to be cited in an AI-generated answer. No provider has published a weighting, a requirement, or a promise. Anyone selling a "get cited by AI" schema package is charging for a correlation nobody has demonstrated.

The reasonable position sits in between. Accurate markup costs nothing extra once the plugin is running, it plausibly helps machines read you correctly, and it's a bad foundation for a content strategy. The things that appear to earn citations — clear direct answers near the top of a page, specific figures, a source worth quoting — are editorial choices, not markup. Whether any of it moves is slow to observe, which is what rank tracking is for.

Run one page through the test tonight

Open the Rich Results Test, paste in your highest-traffic URL, and read the report line by line.

You'll learn three things in about ten minutes: which types that page already qualifies for, whether anything is duplicated, and whether the author, date, and organisation details are still true. If they're stale, fix them in your plugin's settings once — those fields are sitewide, so a single correction propagates to every post you've ever published.

That's the entire job for most small sites. Not adding schema markup, but reading what has been silently published in your name for years and making sure it still describes you.