Why schema is a 1,000-page problem

Open the source of a typical small-business site and you'll often find no structured data at all, or one generic block copied onto every page. Agencies that do it well have something real to sell. The trouble starts when you try to do it for every page of every client.

What is schema, in plain English?

Schema is a short block of code, usually JSON-LD, that sits in a page and states its facts in a format machines can read: this is a business, here is its name, address and phone number; this page is a service; this post was written by this person. Visitors never see it. Search engines use it to understand the page, and Google uses some types to show rich results.¹

The value for AI search is the same idea. Structured data makes a page's facts explicit and unambiguous for any system reading it. It doesn't guarantee a site will be cited in an AI answer, and no honest agency should promise that. It removes guesswork about what the page says.

Why does every page need its own?

Because each page says something different, and the markup has to match it:

• Homepage: Organization or LocalBusiness, built from the client's Business Profile and the homepage itself: name, address, phone, hours, services, area served.
• Each service page: Service markup describing that one service, from that page's own content.
• Pages with a real FAQ section: FAQPage markup that matches the questions and answers on the page.
• Blog posts: Article markup with the author, so the post is tied to a real person.
• E-commerce: Product markup, with reviews and ratings where they genuinely exist.

A site with 40 service pages and 900 posts needs close to a thousand blocks, each one different. Copying the homepage block onto every page describes nothing accurately.

A note on FAQ markup

Where do the hours go?

Writing the markup is only the first step. For each page, someone has to:

1. Work out the page type and which markup it needs.
2. Pull the facts: from the Business Profile for the homepage, from the page itself for everything else.
3. Write the JSON-LD.
4. Run it through a validator and fix what fails.
5. Add it to the page, by hand, on that client's platform.
6. Confirm later in Search Console that it was picked up.

Done one page at a time, a large site turns into weeks of work, and it usually waits until design and development changes are finished. That's why so many agencies sell schema and then do it for the homepage and a handful of service pages.

How does a system do it at scale?

The rules for each page type are written once, in the agency's schema folder: which markup for which kind of page, which fields are required, where each fact comes from. The client's own facts (name, address, phone, service area) live in that client's folder. Then:

1. The system reads each page and decides its type by the agency's rules, flagging anything it isn't sure about.
2. It writes the markup from that page's own content and that client's facts.
3. Every block is checked with Google's Rich Results Test or the Schema Markup Validator.
4. A review page lists each page, its markup, and anything that didn't validate or didn't match the page.
5. A person approves, and only approved markup is added to the site.

When Google changes a rule, the agency updates the schema folder once, and every client's next pass follows it.

What stays with people?

Judgment. Someone decides which pages matter most, whether a page's FAQ is real enough to mark up, and whether a client's reviews meet Google's guidelines. The system does the repetitive part: the hundredth service page, the nine-hundredth post. A person approves every block before it goes live.

Common questions

Does every page need its own schema?

Every customer-facing page benefits from markup that describes what's actually on that page. Copying one block across every page describes nothing accurately.

Does FAQ schema still get rich results in Google?

For most sites, no. Since August 2023 Google shows FAQ rich results only for well-known, authoritative government and health websites. The markup can still describe a real FAQ on the page.

Does schema make a site show up in AI search answers?

No markup guarantees that. Structured data makes a page's facts explicit and machine-readable. It supports good content; it doesn't replace it.

How do you check AI-generated schema?

Every block is validated, compared against the page it describes, and approved by a person before it goes on the site.