Schema markup for small businesses: a practical guide

Schema.org provides a vocabulary; JSON-LD is one way to express it in a page. Google recommends JSON-LD when your setup supports it. Many content management systems and SEO plugins generate it already, so inspect the rendered page before adding a second implementation. A structured-data block should describe the page a visitor can actually read. It is not a place for hidden offers, invented reviews or unrelated search terms.

Choose the entity you actually describe. Organization can describe the business; LocalBusiness can suit an eligible physical business. Product is appropriate for a genuine product page, and Article for editorial content. BreadcrumbList describes a real hierarchy. An online service should not invent an address or pretend to be a local shop. Schema.org supports more types and properties than Google uses for rich results: check Google’s documentation for the particular feature before deciding what to implement.

Open an important page and inspect its existing structured data, including output from your theme and plugins. Assign one implementation to maintain the main entity. Enter consistent names, URLs and other facts. Use Google’s Rich Results Test for supported Google features and Schema.org’s validator for the broader vocabulary. Test a live URL as well as a code sample, because templates, caching and rendering can change what crawlers receive. Fix required-field errors before working on optional enhancements.

Technical validity is only one part of eligibility. Google also applies its content and feature policies and decides whether to show an enhanced result for a particular search. For example, a normal small-business FAQ does not automatically qualify for Google FAQ rich results. Do not promise review stars simply because a page contains review markup. Keep the visible information accurate and check the specific rules for the entity and result type you are using.

Clear information can help machines interpret a page, but it does not guarantee that an assistant will mention the business. Google explicitly says its AI features do not require special AI files or special Schema.org markup. Publish useful explanations and verifiable facts, then treat individual AI-answer tests as samples. Avoid unsupported claims such as a fixed multiplier for AI citations. Read our audit methodology and limitations to separate public website signals from actual business outcomes.

Recheck templates after changing a price, opening hours, product availability, a plugin or a website theme. Compare markup with the visible page and investigate relevant Search Console enhancement reports where available. Save the tested URL, date and owner of the fix. Use the weekly website monitoring checklist for follow-up. A Marketing Scan account can organise website findings; it is not a promise of rich-result eligibility or rankings.