Local business schema: what it earns and what it cannot move

I searched "local business schema" on 6 September 2026 and read the seven results on the first page. Five say the markup improves local rankings, or makes a business more likely to turn up in the local pack and on Google Maps. Not one quotes the sentence in Google's own general structured data guidelines that settles it: a structured data manual action, Google writes, "doesn't affect how the page ranks in Google web search". The query returns an AI Overview built out of the pages that never quote it.

What is local business schema?

Local business schema is JSON-LD markup on your own pages declaring a business as a schema.org LocalBusiness type. Google requires two properties, name and address, and treats it as eligibility for rich results, not as a ranking signal. A single-location site needs one block; a ten-location site needs ten; a service-area business often needs none.

Quick answer

• The markup buys eligibility, not position. Google's guidelines say losing it costs a rich result and nothing else.
• The local pack ranks profiles, not pages. Relevance, distance and prominence are all read off a Business Profile.
• Two properties are required. Name and address. Everything else is optional, and two of the recommended ones do not apply to your own business.
• Google's local ranking page lists five actions, and all five happen inside the profile. The website is not among them.

What kind of problem this is

The model that produces the disappointment treats markup as a lever: something added to a page so that the page moves. Under it, more properties means more signal, and the next question is which properties carry the most weight. That question has no answer, because the premise is wrong structurally rather than by degree.

The numbers travelling with the claim have the same shape. One of the seven pages credits the markup with "a 20-30% increase in click-through rates" and reports that "72.6% of pages ranking on the first page of Google use schema markup", and both are unsourced: the page names no study for either, and the second is a correlation between two things that both track how large a site already is.

Structured data is a set of claims about a page. The local pack does not rank pages. It ranks Business Profiles, which are records inside a different Google product, with their own verification, their own primary category and their own review corpus. A page cannot be a ranking factor in a list that is not made of pages.

The confusion is arrived at honestly, because both places hold the same four facts: name, address, phone number, opening hours, written twice. Two copies of the same data invite the assumption that they feed the same machine, and the assumption survives because nothing ever contradicts it out loud.

The two documents that settle it

Google's general structured data guidelines, opened on 6 September 2026, carry three sentences that between them close the ranking question. The first: "A structured data manual action means that a page loses eligibility for appearance as a rich result; it doesn't affect how the page ranks in Google web search." The second: "Google does not guarantee that your structured data will show up in search results, even if your page is marked up correctly according to the Rich Results Test." And the third, which is the whole economics of this layer in one line: "Using structured data enables a feature to be present, it does not guarantee that it will be present."

The second document is Google's Business Profile help on local ranking. It defines the three factors that decide the map pack, and prominence, the one people assume the website feeds, is "how well-known a business is", based on "how many websites link to your business and how many reviews you have". Links and reviews, not markup. The same page states that "there's no way to request or pay for a better local ranking on Google", and the mechanics of the three factors are unpacked in what local SEO actually ranks.

The countable detail on that page beats the definitions. It lists five things a business can do to improve its local ranking: verify the business, keep the information current, respond to reviews, add photos and videos, and add in-store products. All five happen inside the profile. The website is absent from that list with markup and absent without it, which is why the work that moves the pack sits in how to rank in Google Maps and not here.

What the markup does that the Business Profile does not

Google names four sources for the information in a profile, and the first is "Publicly-available information, such as crawled web content (e.g., information from a business' official website)". The website is not a competitor to the profile and not a shortcut into it. It is one of four inputs Google says it reads, and the markup is what stops that reading from being a guess.

The separation is drawn again on Google's page for adding business details, last updated 10 December 2025. Maps and the knowledge panel go through the profile: "Manage how your business appears on Google Maps and Google Search by claiming your Business Profile." The structured data section of the same page recommends the site logo and breadcrumbs, and never names LocalBusiness. That reads as an omission until you notice the page is routing each job to the surface that owns it. Claiming and verifying the profile, field by field, is covered in optimizing a Google Business Profile.

So the answer to what the markup adds is narrow and real. It makes the page's own version of the facts unambiguous to every consumer that is not the Maps index: Bing, Apple, voice assistants, the data aggregators that license business records, and any crawler reading your HTML rather than querying a Google API. None of them has a verified profile to fall back on. They have your page.

Which makes the disagreement case the interesting one. When the page says one thing and the profile another, the profile wins where the profile is shown, and the markup has quietly published a second version of your address to everything else. The defect is not in the markup, it is that two records drifted, and that same drift appears across directories as inconsistent local SEO citations.

The properties that matter, and the two most guides get wrong

Two properties are required: name and address. That is the entire mandatory surface, which is worth knowing before anyone quotes a fee for implementing it. Google's instruction on the type is to "use the most specific LocalBusiness sub-type possible", so a dental practice declares Dentist rather than LocalBusiness, and a business that genuinely does several things declares an array, because "additionalType isn't supported".

{
  "@context": "https://schema.org",
  "@type": ["Electrician", "Plumber"],
  "name": "The name on the sign outside",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "148 W 51st St",
    "addressLocality": "New York",
    "addressRegion": "NY",
    "postalCode": "10019",
    "addressCountry": "US"
  },
  "telephone": "+12122459600",
  "url": "https://www.example.com/locations/manhattan",
  "geo": {
    "@type": "GeoCoordinates",
    "latitude": 40.761293,
    "longitude": -73.982294
  },
  "openingHoursSpecification": [{
    "@type": "OpeningHoursSpecification",
    "dayOfWeek": ["Monday", "Tuesday"],
    "opens": "09:00",
    "closes": "18:00"
  }]
}

The two properties most guides get wrong are aggregateRating and review. Google's property table marks both with the same qualifier: "This property is only recommended for sites that capture reviews about other local businesses." The review snippet documentation, last updated 24 July 2026, says it harder: "If the entity that's being reviewed controls the reviews about itself, their pages that use LocalBusiness or any other type of Organization structured data are ineligible for star review feature." Four of the seven pages on that results page recommend one or both anyway, on the business's own markup, and one attaches a click-through figure to the practice without a source. Stars for a local business come from the review corpus on the profile, which is why the asking routine in how to get Google reviews is the version of this work that pays.

Two smaller constraints break in silence rather than in the validator. Coordinates need "at least 5 decimal places", so the four-decimal value a mapping tool hands you is accepted and degraded. And priceRange "must be shorter than 100 characters. If it's 100 characters or longer, Google won't show a price range for the business", which is how a paragraph pasted into that field disappears with no error anywhere.

Retired features deserve their own line, because location pages are still being built with them. FAQ rich results stopped appearing on 7 May 2026, and Google removed the documentation on 15 June 2026, noting the feature "is no longer shown in Google Search results". The How-to rich result went the same way on 14 September 2023, "on both desktop and mobile devices". FAQPage markup on a location page today is valid, harmless and produces nothing visible. The same holds for markup added in the hope of an AI answer: Google's generative AI guide files structured data under what to ignore, writing that "structured data isn't required for generative AI search, and there's no special schema.org markup you need to add", a position developed further in how to rank in AI Overviews.

When NOT to spend time on this

When the Business Profile is not verified yet. Every hour spent on markup is an hour the record that actually ranks went without one, and the sequence is not a matter of taste.

When the page is not indexed. Markup on a page Google has not stored is a claim nobody received, and the diagnosis for that belongs in why a page is not indexed.

When the platform already emits it. Most modern site builders and local SEO plugins output a LocalBusiness block by default. Adding a second one produces two entities where the page describes one.

When the facts are not on the page. Marking up an address the visitor cannot see violates an explicit guideline, and for a service-area business that hides its address, the correct amount of address markup is none.

What changes with the shape of the business

One location. The block goes on the page a customer would land on, usually the contact or the home page, and it takes about twenty minutes once the facts are settled. The eligibility it buys is a knowledge panel, and the honest expectation is that most days nothing observable happens.

Several locations. This is the case the markup was built for, and the only one where it earns real work. One block per location page, each with its own address and coordinates, generated from the same source that renders the visible page rather than typed. The failure mode at scale is not invalid JSON, it is drift: a template that keeps emitting last year's hours for forty branches, valid the whole time, wrong the whole time. The wider architecture question underneath that is handled in SEO for multiple locations.

Service area, no public address. A plumber who works out of a van has a profile with the address hidden and a page with no address on it. There is nothing to mark up beyond name, phone and area served, and the guides that hand out a full template push these businesses toward publishing an address they deliberately hid.

Mistakes that repeat

• Expecting positions. The one metric it cannot move is the one people watch afterwards, so the layer gets written off as useless.
• Marking up your own star ratings. Google names LocalBusiness specifically as ineligible when the entity controls the reviews.
• Copying a template with every property filled. Declaring facts that are not on the page is a guideline violation, not a completeness score.
• Letting the markup and the profile drift apart. Two addresses in circulation is worse than one address stated once.
• Reading the Rich Results Test as a validator. It reports eligibility for a feature, so a correct block for a retired feature reads like an error.
• Adding a second block on top of the platform's. One page, one business, one entity, and the plugin got there first.

Data and transparency

The three quoted sentences on eligibility and ranking come from Google's general structured data guidelines. The required properties, the sub-type instruction, the additionalType note, the five-decimal coordinate rule and the 100-character priceRange limit come from Google's Local Business structured data page, last updated 10 December 2025. The self-review restriction comes from Google's review snippet documentation, last updated 24 July 2026. The retirement dates for FAQ and How-to rich results, and the sentence about structured data not being required for generative AI features, come from Google's Search Central changelog and its AI features guide. All were opened on 6 September 2026.

The three local ranking factors, the five suggested actions, the definition of prominence and the line that there is no way to pay for a better local ranking come from Google's Business Profile help on local ranking, and the four sources Google names for profile information come from its page on how it sources that information, both opened on 6 September 2026. The counts of what the first page of results claims, five of seven asserting a ranking or local pack effect, four of seven recommending self-applied rating markup and none of seven quoting the ranking sentence, are my own reading of that results page on that date, repeatable by anyone in fifteen minutes. The two figures quoted from one of those pages, a 20-30% click-through increase and 72.6% of first-page results carrying markup, were copied verbatim from it that day and appear here as examples of numbers arriving without a source, since that page names no study for either; they are not reproduced as measurements. No percentage of local ranking weight appears in this article because Google publishes none. Judgment about implementation order comes from operating sites that record more than 300 million impressions a year in Search Console. Verified as of September 2026.

What this changes

The uncomfortable part is who benefits from the confusion. An owner told that schema markup improves local rankings will pay for it, see no movement, and write the whole layer off as worthless, when the layer did exactly what the documentation said and the promise came from somebody else. The markup is cheap, bounded and correct to have. It is also the only piece of local search work that can be delivered, invoiced and screenshotted without anyone logging into the profile where the ranking is decided, and that convenience explains a great deal about which of the two gets sold. The test before approving a quote is short: ask what the deliverable makes the business eligible for, and if the answer names a position instead of a feature, the quote describes a different product than the one being built. Then check whether the items in the local SEO checklist that do move the pack were ever touched.

Frequently asked questions