The record that tells Google what your business actually is

Your website can carry a structured record of your business facts — what you are, where, when you're open, what you charge, who runs it — written in a format machines read directly. Without it, search engines infer those facts from your page text. Inference is where wrong phone numbers and wrong holiday hours come from.

The industry calls this schema, or JSON-LD. We call it your machine fact sheet, because that's what it is.

What difference does it actually make?

Two things. First, accuracy: you're telling rather than hoping. Second, eligibility — the enhanced results that make one listing visibly bigger than the six around it. Star ratings, prices, opening hours, expandable questions. You can't get those without a record, and having one doesn't guarantee them either. It makes you eligible; Google decides.

A concrete version: a restaurant that states its holiday hours in machine-readable form has correct hours on the Google panel on Thanksgiving. One that only wrote them in a paragraph has a customer at a locked door, and a one-star review about it.

What goes in it?

• Your business, using the most specific correct category — not "business", but the actual thing you are
• Address broken into its proper parts, map coordinates, phone, email, hours including holidays, the area you serve
• Each service, with what it is and what it costs
• Your articles, with truthful publication and update dates and a real named author
• Your Q&A sections — generated from the same source as the visible questions, so the two can never drift apart
• Every profile that proves you exist elsewhere, linked back

The pieces are then joined by internal references, so a machine reads one business, which has a website, services, articles and a founder — rather than a pile of unrelated facts that happen to share a page.

What's the most common way this goes wrong?

Drift. The visible page says one thing and the record says another, usually because someone edited the page and not the record. The second most common is a syntax error: one stray comma silently deletes the entire record while the page still looks perfect to you.

That's why the check that matters is done on the file a visitor actually receives, not on the source and not on a staging copy. If nobody parsed the shipped page, nobody knows whether the record survived the build.

Can I just claim good things in it?

No, and this is worth being blunt about. Only facts your site genuinely states belong in the record. Price range, founding date, founder, ratings — these are facts, not blanks to fill with something plausible.

Review and rating markup in particular: it only goes in when there are genuine first-party reviews shown on that page. Fabricating it is a manual-penalty risk, and more to the point it's a lie told in a format specifically designed to be trusted.

Common questions

Is this the same as llms.txt?

No. llms.txt is a plain-text summary file for AI systems — worth publishing, costs nothing, and honestly it's future-proofing rather than something driving visibility today. The structured record is the one doing real work now.

Will this get me star ratings in Google?

It makes you eligible. Google decides whether to show them, and it changes its mind about entire categories periodically without notice. Anyone promising the stars themselves is promising something they don't control.

My site is on a platform that adds schema automatically. Am I covered?

Partly. Automatic markup tends to be generic and often asserts things your pages don't state. It's a starting point, not a finished job — and worth checking, because a platform update silently deleting it is one of the failures we look for every month.