· 2 min read

Schema markup for beginners: which types matter for a small site

Most schema advice lists thirty types you'll never need. A small site needs about four. Which ones, what they earn you in 2026, and how to add them safely.

Schema markup (structured data) is machine-readable labeling: a block of JSON-LD in your page that says "this is an article, published on this date, by this organization." Search engines use it to build rich results; AI assistants use it to understand what they're reading before they decide whether to cite it.

It's also where small sites waste the most effort, because most guides list every type that exists instead of the four that matter.

What does schema markup actually earn you in 2026?

Be clear-eyed about this, because the landscape changed:

  • Breadcrumbs — still a live rich result. Your URL line in Google becomes a readable trail. Cheap to add, visible payoff.
  • Product, Review, Recipe, Event, JobPosting — still live and valuable if you genuinely have those things.
  • Article/BlogPosting — no dramatic visual treatment, but it cleanly declares dates and authorship, which matters for freshness signals and for AI extraction.
  • FAQ rich results — effectively dead for normal sites since 2023 (Google limits them to government and health sites). FAQPage markup still helps AI assistants parse your Q&A; just don't expect SERP accordions.
  • HowTo — rich result retired entirely. Same caveat: the markup still structures your content for machines.

The quiet second audience is AI search. Assistants deciding what to cite lean on unambiguous machine-readable claims about what a page is. Schema is how you make those claims.

Which types should a small site actually add?

In order:

  1. Organization — homepage only. Name, URL, logo, and sameAs links to your real profiles. This is identity: it helps machines connect your site to your brand everywhere else it appears.
  2. BreadcrumbList — every interior page. The one with a visible SERP payoff.
  3. Article or BlogPosting — every post, with honest datePublished and dateModified.
  4. Your money type — Product with offers, LocalBusiness with address, Service — whichever one matches what the site sells. Only mark up what's visibly on the page.

That's the whole list for most sites. WebSite, AboutPage, and the exotic types add bytes, not results.

What are the rules for doing it safely?

Markup must describe what's on the page. Google's structured data penalties (they're manual actions — a human reviews your site) come almost entirely from markup that misrepresents: review stars for reviews that don't exist, prices for products you can't buy, FAQ markup on content that isn't Q&A. If a visitor can't see it, don't mark it up.

One broken comma kills the block. JSON-LD that doesn't parse is silently ignored — no error, no rich result, nothing. This is the most common schema failure and the easiest to catch: run your page through our schema markup validator, which reports parse errors per block and missing required properties per type.

Required properties gate the rich result. A Product without offers, an Article without datePublished — valid JSON, no payoff. The validator flags these too.

How do I add schema without a developer?

Most CMSs have made this a settings toggle: WordPress SEO plugins emit Article and Breadcrumb markup automatically; Shopify themes include Product markup; site builders vary. So the practical workflow is:

  1. Check what you already have — run your homepage and one post through the validator. You may be done already.
  2. Fill gaps at the template level, not page by page.
  3. Re-validate after any theme or plugin change — schema regressions ship silently.

This checked one page. Your site has more of them.

RankRocket crawls your entire site, scores every page with the same rubric, finds duplicate titles and descriptions across pages, and turns it all into a short daily fix list. Free — no card required.

Audit my whole site free