Schema Markup for Service Businesses: A 2025 Implementation Guide

Schema markup is structured code embedded in a webpage that tells engines what the content means rather than asking them to infer it from raw text. With clean schema, Google understands that 'John Smith' is a Person, not a city; that the page is about a Service, not a blog post; that '4.9' is an aggregateRating, not a price. Pages with proper schema win rich results in classic SERPs, are more likely to be cited by AI engines, and are easier for voice assistants to read aloud.

For a service business, schema also closes the gap between what you do and how you describe it. The website may say 'we're a friendly family-owned plumbing company' but until the LocalBusiness schema declares the business name, address, phone, hours, areaServed, services, and aggregateRating, the engines treat that as marketing copy rather than structured fact.

LocalBusiness on the homepage. The most important entity on the site. Include name, address, phone, openingHours, areaServed, priceRange, aggregateRating, paymentAccepted, sameAs (social profiles), and a logo. For sub-types (Plumber, Dentist, LawFirm, MedicalBusiness), use the most specific available type from schema.org.

Service per named service. One Service node per top-level service offered, each with name, description, provider (linked by @id to the LocalBusiness), areaServed, hasOfferCatalog where appropriate. This is what Google references when it surfaces a service in the GBP services tab and what AI engines lean on for category-specific queries.

FAQPage on every page with FAQs. Mark up each question and answer pair. Five to ten FAQs per cornerstone page is the sweet spot — enough to capture variant queries, not so many that signal-to-noise drops.

BreadcrumbList on every interior page. Tells engines the hierarchy and helps the site-link breadcrumb display in classic SERPs. Cheap to add, consistently rewarded.

Article on every blog post. Include headline, datePublished, dateModified, author (linked by @id to a Person node), publisher (linked by @id to the Organization), image, and articleSection. AI engines lean on Article schema heavily when deciding whether to cite a blog post.

JSON-LD is the format Google recommends. Embed in a `<script type="application/ld+json">` block in the head or body. For sites with multiple entities per page (which is most pages), use a single @graph block that links all entities by @id rather than emitting many disconnected JSON-LD blocks.

Linking by @id matters because it tells the model these are the same entity rather than coincidentally similar. The Person 'Taylor Moses' on the about page and on the post byline should reference the same @id. The Organization on the homepage and the publisher of the article should reference the same @id. Engines reward linked entities with stronger entity recognition and richer rich-results.

Validate everything. Google's Rich Results Test catches errors and warnings; the Schema.org validator catches structural issues. Both should run as part of every release. Broken schema is worse than missing schema because the engines may distrust the page entirely.

Mismatched on-page text and schema. The schema says aggregateRating 4.9 but the page shows no review section. Google penalizes this aggressively as fake markup. Every fact in schema must be visible on the page.

Forgetting dateModified. Articles without a recent dateModified look stale to AI engines. Update dateModified whenever the content meaningfully changes.

Generic Person nodes for authors. Author entities should include name, jobTitle, image, sameAs (LinkedIn, Twitter), and a knowsAbout array. Bare Person nodes signal less authority.

FAQPage stuffing. More than 15 FAQs per page often gets ignored by Google or triggers a manual action. Five to ten high-quality FAQs win.

Schema in the wrong format. Microdata and RDFa still work but JSON-LD is preferred. Mixing formats on the same page confuses crawlers.

Audit first. Crawl the site with Screaming Frog or Sitebulb and export current schema. Identify gaps (missing types) and errors (invalid markup). Most service-business sites we audit have FAQPage schema in 0% of relevant pages and LocalBusiness schema with three required fields missing.

Implement page-by-page in priority order: homepage (LocalBusiness), services pages (Service + FAQPage), location pages (LocalBusiness with areaServed), then blog posts (Article). Validate each page before moving to the next.

Track citation and rich-result impressions in Google Search Console. Within 4–6 weeks, FAQ rich results, sitelink expansions, and AI-Overview citations should grow. Track monthly.