LocalBusiness vs Restaurant Schema: Which for Which Page?

Picking a Schema.org type looks trivial, but the wrong choice can hide you from AI engines for months. A Kadıköy venue used generic LocalBusiness on its homepage and was excluded from restaurant-specific Perplexity answers for six months. Switching to Restaurant surfaced it in 11 new queries.

The Hierarchy: Restaurant Is Already LocalBusiness

In Schema.org the chain reads Thing → Place → LocalBusiness → FoodEstablishment → Restaurant. When you declare @type Restaurant you inherit every LocalBusiness property (address, telephone, openingHours, geo, aggregateRating).

Using the generic LocalBusiness type makes AI agents skip you on food-vertical prompts such as "restaurants nearby" or "breakfast in Kadıköy". Engines apply a type filter: anything that is not @type Restaurant is not a candidate for food answers.

Page-by-Page Decision Tree

Match the schema type to the page purpose. One type does not fit all surfaces:

• Homepage: Restaurant (servesCuisine, menu, priceRange, hasMenu)
• Blog post: Article or BlogPosting + author (Person) + datePublished
• Campaign / event page: Event (startDate, endDate, location, offers)

Three Frequent Mistakes

The first is embedding multiple LocalBusiness types side by side on the homepage. Two peers confuse the AI parser. Use @graph with Restaurant as the main node and the rest as references.

The second is reusing the Restaurant type inside blog posts. AI then reads the post as a new venue, not editorial content; Article is mandatory. The third is wrapping campaigns in plain Offer markup. Event + offers is a stronger trigger for queries like "deals in Kadıköy this week".

FAQ

Can I declare both LocalBusiness and Restaurant? Yes (@type: ["Restaurant", "LocalBusiness"]) is valid but redundant — Restaurant already implies LocalBusiness. Keep it single.

How do I distinguish a cafe from a restaurant? Use CafeOrCoffeeShop when coffee is the core service; Restaurant when you serve a full menu.

How long until AI engines reflect a schema change? Perplexity and ChatGPT recrawl in 3-6 weeks; visible impact typically lands in week 4 and plateaus around month 2-3.

Found this helpful? Share it.