Schema markup for owners who don't read JSON

Every consultant in 2026 mentions schema markup as though the business owner is supposed to know what it is. Most don’t, and the consultants don’t explain it because explaining it isn’t what they’re hired for. The result is that owners commit budget to a thing they can’t evaluate.

I want to fix that. This essay is the plain-English version. No code, no JSON, no examples that assume you’ve read previous documentation. Just what schema is, why it matters more this year than last, and what to do about it without reading a single line of markup.

What schema actually is

A website page has two layers. The visible layer is what a human sees — text, images, formatting. The invisible layer is what a machine reads — code that describes the page’s structure and contents in a standardized way.

Schema is the invisible layer’s vocabulary. When a search engine or AI assistant fetches your page, it reads both layers. From the visible layer it gets content. From the schema it gets meaning. Schema is how you tell a machine, formally, this paragraph is a customer review; this number is the price; this is the business’s phone number; these are the hours we’re open.

That’s it. There’s no magic. The machine could try to figure all of this out by reading the visible page and guessing — and sometimes it does. But guessing is unreliable, and a page with explicit schema is dramatically more accurately understood than a page without.

Why it matters more in 2026 than it did in 2022

Two compounding reasons.

The first is that Google’s ranking systems have grown more dependent on schema for entity recognition. A LocalBusiness schema with address, hours, services, and reviews gives Google a clean handle on the business. Without it, Google has to do extra work to figure out what kind of business you are and where you operate. Google’s ranking systems reward the cleaner handle. They demote the page that makes them guess.

The second reason — the one that’s new in the last two years — is that generative AI assistants depend on schema even more than Google does. When ChatGPT or Perplexity browses your page during a conversation, the model has limited time to extract a usable answer. Schema is the cheat sheet that lets the model lift specific facts cleanly. A page with FAQPage schema gives the model a structured question-and-answer pair it can drop into a response. A page without it gives the model a wall of text to summarize on the fly, which the model often does badly or skips entirely.

The brands being cited inside AI Overviews and ChatGPT in 2026 are disproportionately the brands with rich schema. The brands being skipped are the ones who never added it.

What kinds of schema matter for a typical business

There are dozens of schema types in the formal Schema.org vocabulary. For a service business the relevant subset is small.

LocalBusiness or its more specific subtype — ProfessionalService for consultancies, MedicalBusiness for clinics, Restaurant for restaurants, and so on. This describes the business as an entity: name, address, hours, telephone, services. It’s the most foundational piece of schema a service business can have, and most don’t have it.

Service — describes each individual service the business offers. We provide tax preparation. We provide audit response. We provide bookkeeping. Each service is its own structured entity. Three or four Service entries inside the LocalBusiness schema is normal.

FAQPage — describes question-and-answer sections on a page. If your page has a “frequently asked questions” block, marking it as FAQPage schema lets AI assistants lift the Q&A pairs directly into their responses. This is the highest-leverage schema type for AI search right now.

Review and AggregateRating — describes customer reviews and the average rating. Most businesses display reviews visually but never mark them as schema. The ones who do get the gold-star treatment in AI Overviews; the ones who don’t see their competitors get it instead.

Article — for blog posts and journal entries. Names the author, publication date, headline, body. This is what lets a model attribute a piece of content correctly when it cites it.

Five types. That’s the entire kit for a typical service business.

What to do without reading the markup

The owner’s job isn’t to write schema. It’s to make sure schema is present, accurate, and validated.

Three concrete moves:

First, go to validator.schema.org and paste in your homepage URL. The tool will tell you what schema is present and whether it’s valid. Most owners doing this for the first time find that they have either zero schema, or invalid schema, or schema that names a different business than the one they actually own (the agency built the site for a template client and never updated the schema). Twenty minutes of audit, before any work begins.

Second, ask whoever maintains your website — agency, freelancer, in-house — to add the five schema types above. The work, for a typical site, is one to two hours. If your provider doesn’t know what schema is, find a different provider. This is now non-optional for a 2026 website.

Third, audit it again in a quarter. Schema can break when sites get redesigned, when CMS plugins update, when content moves. A page that had valid schema in March can quietly lose it by July. The quarterly check takes ten minutes and protects an asset that took hours to build.

The single sentence to take away

Schema is the difference between a search engine understanding your business cleanly and guessing about it badly — and in 2026, the brands being guessed about are the ones being cited less, ranked lower, and recommended to fewer buyers, every quarter.

That’s the whole pitch. You don’t need to read the code. You need to know it’s there, valid, and current.

---

Luke LaFave is the founder of LaFave Consulting. The studio’s custom website builds include hand-authored, validated schema across all of the types described above.