Schema Markup: The Complete Guide

Schema markup is a standardized code format, usually written as JSON-LD, that tells search engines and AI crawlers exactly what a webpage contains rather than leaving them to guess from plain text. It sits invisibly in a page's code and does not change what visitors see, but it gives machines a structured, unambiguous description of the content underneath.

What Schema Markup Actually Does

Search engines and AI systems are constantly trying to understand what a page is about. Plain text and headings give clues, but they leave room for ambiguity. A page might mention a product, a price, and a review, but without structured data, a crawler has to infer how those pieces relate to each other. Schema markup removes that guesswork by explicitly labeling each piece of information: this is a product, this is its price, this is a review of it, and this is the rating given.

This matters for two connected reasons. First, structured data can unlock rich results in traditional Google search, star ratings under a product listing, an event date and location, or a recipe's cook time displayed directly in the results. Second, and increasingly important, it gives AI systems like ChatGPT, Perplexity, and Google's AI Overview a clean, structured source of facts to draw from when generating an answer, which matters more each year as AI mediated search grows.

There is a useful mental model for understanding why this matters. A human reader can infer meaning from context, layout, and design cues, a bold heading followed by a price clearly reads as a product listing to a person. Machines cannot rely on those same visual cues nearly as reliably. Schema markup translates that same information into an explicit, unambiguous format a machine can parse with certainty rather than inference.

Where Schema Markup Comes From

Schema markup follows a shared vocabulary defined at schema.org, a collaborative project originally backed by Google, Microsoft, Yahoo, and Yandex. Google implements and documents its own supported subset of this vocabulary through its own structured data guidelines, which explain exactly how markup should be formatted and which types are eligible for specific search features. Not every schema.org type produces a visible search feature in Google, but a type does not need to unlock a rich result to still be useful for general machine readability and AI citation.

JSON-LD Is The Recommended Format

Structured data can technically be written in a few different formats, but Google explicitly recommends JSON-LD, a small block of JavaScript style code placed in a page's head section, separate from the visible HTML content. This separation makes JSON-LD easier to implement and maintain than older formats that required embedding markup directly inside visible page elements, since a mistake in the visible content will not accidentally break the structured data, and vice versa.

The Schema Types That Matter Most For Most Businesses

Organization Schema

This defines core facts about a business as an entity: its name, logo, official website, and links to verified social profiles through the sameAs property. Organization schema is typically the foundation of a site's structured data strategy, since it establishes the entity that other schema types reference.

LocalBusiness Schema

An extension of Organization schema built specifically for businesses with a physical location or defined service area. It adds fields for address, phone number, business hours, and geographic coordinates, directly reinforcing the same NAP consistency signals that matter for local search rankings.

Product Schema

Used on ecommerce and product pages to define price, availability, and review information. This is one of the most visually impactful schema types in traditional search, often unlocking star ratings and price display directly in search results.

Article Schema

Applied to blog posts and news content, defining the headline, publish date, author, and publisher. This helps establish authorship and freshness signals that matter for both traditional SEO and the kind of authority signals AI systems weigh when deciding what to cite.

FAQPage Schema

Marks up genuine question and answer content written by the site owner. While Google removed the visual FAQ rich result from search in 2026, the underlying schema remains valid and continues to support machine readability and AI citation, a distinction covered in more depth in our post on why FAQ schema is not dead, it just changed jobs.

BreadcrumbList Schema

Defines a page's position within a site's navigation hierarchy, helping search engines understand site structure and often producing a breadcrumb trail directly in search results instead of a raw URL.

Review Schema

Marks up individual reviews or aggregate rating data. This type carries stricter eligibility requirements from Google than most other types, since review content has historically been a target for manipulation, so implementation needs to closely follow Google's guidelines to avoid the markup being ignored or penalized.

How To Implement Schema Markup Correctly

• Match the markup to what is actually visible on the page. Structured data should accurately reflect visible content. Marking up a five star rating that does not actually appear anywhere on the page, for example, violates Google's guidelines and can result in the markup being ignored or the page being penalized.
• Use the most specific type available. Schema.org offers general types alongside more specific ones. A furniture retailer's product page should generally use a specific type over a generic fallback, since precision strengthens the relevance signal for both search engines and AI systems.
• Validate before publishing. Google's own Rich Results Test and Schema.org's general validator both check whether markup is technically correct and eligible for specific search features. Testing before publishing catches formatting errors that would otherwise silently prevent the markup from being read at all.
• Keep markup updated as content changes. If a product's price changes, or an article gets a substantial edit, the corresponding schema needs to be updated too. Stale structured data that no longer matches the visible page is treated the same as any other mismatch between markup and content.
• Do not over mark up a page. Applying every possible schema type to a single page, regardless of relevance, does not provide extra benefit and can create internal contradictions if types overlap incorrectly. A page should carry the schema types that genuinely describe its content, not the maximum number available.

Schema Markup Versus Meta Tags: They Are Not The Same Thing

These two are frequently confused, so it is worth being precise. Meta tags, including the title tag and meta description, are a much older HTML standard that controls what appears in a search result snippet and browser tab. Schema markup is a separate, newer layer that describes the actual content and entities on a page in a structured, machine readable way. A page can have excellent meta tags and no schema markup at all, or the reverse, since the two serve different functions. Strong technical SEO uses both together rather than treating one as a substitute for the other.

Schema Markup For Multi Location And Service Area Businesses

Businesses operating across several cities face a specific structured data challenge worth addressing directly. Each location typically needs its own LocalBusiness schema block, matched to its own dedicated page, rather than a single Organization schema attempting to represent an entire multi city footprint. Using the parentOrganization property to link individual location schema blocks back to a central Organization entity keeps the entity relationships clear for search engines and AI systems alike, mirroring the same structure Google expects across a business's actual Google Business Profile listings for each location.

Why Schema Markup Matters More In The Age Of AI Search

Traditional SEO valued schema markup primarily for the visual search features it could unlock. That value has not disappeared, but a second, arguably larger value has emerged alongside it. Generative engines like ChatGPT, Perplexity, and Google's AI Overview rely heavily on clearly structured, unambiguous content when deciding what to cite in a generated answer. A page with clean, accurate schema markup gives these systems exactly the kind of machine readable facts they are built to extract, reducing the chance that an AI system misinterprets or skips the content entirely in favor of a competitor's more clearly structured page. This overlap between traditional structured data best practice and AI visibility is explored further in our guide to AEO vs GEO vs SEO.

Common Schema Markup Mistakes

• Copying a competitor's schema without adapting it. Structured data must match your own page's actual content. Reusing another site's markup wholesale, rather than adapting it to reflect what is genuinely on your page, is a fast path to invalid or misleading data.
• Ignoring validation errors. A single formatting mistake, like a missing comma or an incorrectly nested property, can invalidate an entire JSON-LD block. Testing is not optional, since broken markup is often silently ignored rather than flagged anywhere visible to a site owner.
• Assuming schema markup guarantees a rich result. Correct markup makes a page eligible for certain search features, but eligibility is not the same as guaranteed display. Google ultimately decides whether and when to show a rich result, based on additional quality signals beyond the markup itself.
• Ignoring how content updates ripple through connected schema types. Since schema types frequently reference each other, updating a Product's price without checking whether an associated Offer or aggregate Review block also needs adjustment can leave a page with internally inconsistent structured data, which is treated the same as any other inaccurate markup.
• Forgetting to update markup after a redesign. Website redesigns and CMS migrations frequently drop or break existing structured data. A post launch audit of schema markup should be standard practice after any significant site change.
• Nesting properties incorrectly. Schema types often reference other schema types, a Product referencing an aggregate Review, for example. Getting this nesting structure wrong is one of the most common technical errors, and it usually causes the entire block to fail validation rather than just the incorrect portion.

Frequently Asked Questions