GEO — AI SearchJSON-LD code snippet of type Restaurant shown next to an enriched Google search result

Structured Data for Restaurants: 2026 Guide

Updated on 6 September 2026

In short

  • Structured data (schema.org, JSON-LD format) helps Google and generative AI understand unambiguously who you are, what you offer, and at what price.
  • It guarantees no ranking — it's a clarity tool, not a ranking lever.
  • The FAQPage schema no longer produces a visual rich result in Google as of 7 May 2026, but remains valid and useful for content understanding.
  • The most useful schemas for a restaurant: Restaurant (or LocalBusiness), Review/AggregateRating, Menu, and Organization for software like SEOresto.

What structured data actually does

Schema.org markup, inserted as JSON-LD in a page's code, explicitly tells a machine what it's looking at: "this is a restaurant, its name is X, it serves Y cuisine, its average rating is Z out of N reviews." Without this markup, Google and generative AI have to infer this information from visible text — which works, but with more uncertainty.

Two distinct effects, not to be confused:

  1. Visual rich results in Google Search (stars, price, availability) — depend on the schema type and change over time.
  2. Entity understanding by search engines and generative AI — a quieter effect, but a permanent one as long as the markup stays in place.

Relevant schemas for a restaurant

SchemaUseCurrent visual effect
Restaurant (subtype of LocalBusiness)Name, address, hours, cuisine type, price rangeCan display hours and rating in some contexts
AggregateRatingAverage rating and review countStar display in some results
ReviewIndividual reviews with author and ratingSupports AggregateRating
Menu / MenuItemDishes, prices, allergens"Menu" rich result depending on Google availability
BreadcrumbListBreadcrumb trailPath display in the result URL
FAQPageQ&A pairsRemoved from rich results on 7 May 2026 — still valid for content understanding

The FAQPage case: what changed

Google removed the enriched Q&A display generated from FAQPage markup from its search results on 7 May 2026. The markup remains technically valid and continues to be parsed by Google to understand page content — but it no longer produces the visual effect (expandable questions) that once justified its blanket use. Practical consequence: only add this schema to pages where a real FAQ section exists, not as an automatic checkbox on every article.

For software like SEOresto

The SoftwareApplication schema, paired with offers and an explicit price, helps generative AI return an exact rate rather than a rough or outdated estimate — a sensitive point when price is a differentiation argument. Organization reinforces entity consistency (name, description, logo) across every page of the site.

How to check the markup works

Google's "Rich Results Test" tool (publicly available at search.google.com) validates markup syntax and shows which rich results are eligible. It's a free check to run after every deployment that touches a page carrying a schema.

A common mistake to avoid

Marking up information that doesn't visibly appear on the page (for example a rating or price present only in the JSON-LD, absent from visible text) violates Google's structured data policies and can trigger a manual action. Markup should always reflect content a human visitor can actually see.


For the overview of generative AI visibility, see get your restaurant recommended by ChatGPT.

SEOresto structures your listing and review data correctly from the start. From €39/month. See pricing.


Sources: schema.org (Restaurant, Review, AggregateRating, FAQPage type specifications); Google Search Central documentation on rich results and the FAQPage removal (May 2026).

Author: Dmitrii Portnov, founder of SEOresto.

Lire cet article en français →

Related articles

← Back to blog