How Google chooses the site name on a result

A site name is the label Google shows for the source of a result. It belongs to the site. The title link belongs to the page. Google’s site names documentation, updated 10 December 2025, says the two are different on purpose: title links are specific to each web page, and the site name is for the whole site. WebSite structured data on the home page is how you state a preference. Google’s system is automatic. It can decline the preference.

What a site name is

When Google lists a page, it can show the name of the site the page comes from. That label is available in the languages where Google Search is available, on mobile and on desktop, for domain-level sites and for subdomain-level sites. It is not the blue title. A reader uses the site name to see who published the result, and the title to see what that one page promises. Adding the brand at the end of every title does not set the site name. Google may already show the site name beside the result, which is why a title that is only the brand, or mostly the brand, wastes the title.

The goal, in Google’s words, is to represent and describe the source of each result. Generation is completely automatic. It uses content from the home page and references to the site that appear on the web. You indicate a preference. You do not type the label into Search Console as a locked field. There is no switch that forces the string onto every URL the same day.

How Google chooses it

WebSite structured data is the most important place to specify the preference, if you want one. The system also considers og:site_name, the title element, heading elements, and other text on the home page. Those sources should tell the same story. A home page whose title says one brand, whose visible heading says another, and whose schema says a keyword is three candidates. Google will pick. Consistency is how you stop the pick from being a surprise.

If the system is less confident in the name you provided, it may generate a name from other sources or show the domain or subdomain. That is a normal outcome, not a broken tag. The work is to make the preferred name obvious and available, then wait for a recrawl. Internal pages can lag the home page. Seeing the right name on the home result and the domain on an article is a processing delay, not proof that articles need their own WebSite block.

How to choose the words

Choose a unique name that reflects the site and is not misleading. It has to follow Search content policies. Use a concise name people already recognize. Google’s example of the shape is “Google,” not “Google, Inc.” There is no character limit, and long names may be truncated on some devices. A legal suffix does not help a person scanning results. The short public name does.

Avoid a generic name. “Best Dentists In Iowa” is unlikely to be selected unless that phrase is an extremely well-recognized brand. A publishing site named after the query it wants is in that pattern. Northwind is a name. “Best Async Standup Blog” is a title wearing a site-name field. If the only reason for the words is a keyword, they belong in a page title when they truly describe that page, not in the source label for every URL.

Use the name the same way everywhere the home page talks about itself: the visible title, the heading, og:site_name if you set it, and the schema. The schema is not a place to hide a second brand. If you already have a WebSite node for something else, nest the name properties in that same node. A second WebSite block on the home page is what Google says to avoid.

Where WebSite schema goes

Google Search supports one site name per site. A site, for this feature, is a domain or a subdomain. It does not support site names at the subdirectory level. https://example.com/ and https://www.example.com/ are domain-level home pages. A host that starts with m. is generally treated as equivalent to those, not as a separate brand. https://news.example.com/ is a subdomain and can have its own name. https://example.com/news is a folder. Markup there does not create a site name for the folder.

The markup goes on the home page only: the root of the domain or of the subdomain. https://example.com/de/index.html is not that home page. You do not repeat the block on every article. Article schema describes the article. The article schema guide is that object. Mixing a WebSite name into every post does not give the post its own source label, and it can disagree with the home page.

If a subdomain’s home page has no WebSite markup, Google may use the domain-level site name as a fallback. That is a reason to add markup on a subdomain you actually want named differently, and a reason not to invent a subdomain just to get a second name. The home page has to be crawlable. If robots.txt, a noindex, or a login blocks it, Google may not be able to generate a site name. The robots meta guide is why the public home page should not be noindex.

name, url, and alternateName

The required properties are name and url. name is the site name, under the choosing rules above. url is the canonical home page of that domain or subdomain, such as https://example.com/ or https://news.example.com/. JSON-LD, RDFa, or microdata are all acceptable. JSON-LD in the home page response is the form that is easiest to keep next to the visible name. Do not depend on a script to invent the name after load if you can write it in the HTML. The JavaScript SEO guide is why the response should already contain the markup.

alternateName is optional. Use it for a real acronym or a shorter name people also use. You can list more than one, in the order you prefer, most important first. Google’s system tries the preferred name and may use an alternate when the preferred name is not available. One reason it would be unavailable: the system generally will not use the same site name for two different sites that are global in nature. Another: the site may be better known by the acronym than by the long name. An alternate is a backup identity, not a list of keywords.

A domain or subdomain can itself be an alternate, and it must be all lowercase for Google to treat it as that kind of preference. example.com can be detected. Example.com is not the form they describe. Put it last in the alternate list if you want it only when the real names are not selected. Using the bare domain as the preferred name is a last resort. Google says the system will generally select it then, and recommends that only when the other steps have failed. A site that is already known by its name should not demote itself to the hostname on day one.

Duplicate home pages and redirects

If the home page exists at more than one URL, such as HTTP and HTTPS, or www and the bare host, put the same WebSite markup on each duplicate, not only on the canonical copy. Google’s note is explicit. A canonical tag does not excuse the other copy from carrying the same name. The canonical guide is still how you pick which home URL is preferred. The site-name rule is that the duplicates agree while that preference is being resolved.

Redirects have to work, and Googlebot has to be able to fetch the target. If the home page redirects, the site name reflects the target. A chain that ends on a URL the crawler cannot read leaves the system without the page you think is home. One hop, to the canonical home, is the redirects pattern. Then the WebSite url is that target, and the name on that target is the name.

HTTP and HTTPS should not disagree. A name on the HTTPS home page and a different name on the HTTP home page is two preferences for one site. Pick the public name once. Copy it. Retire the HTTP URL with a redirect when you can, and until that redirect is the only path, keep the markup aligned.

How to test it

Site names are not supported in the Rich Results Test. A pass or a blank result there does not tell you the name is valid. Use a schema validator to check syntax. Use URL Inspection to see whether Google can fetch the home page, and confirm it is not blocked by robots.txt, noindex, or a login. If the page looks right, you can ask for a recrawl. Google says it can take anywhere from a few days to a few weeks after publishing for a page to be found and crawled, and the same window applies after you change the name. Internal pages need their own time after the home page is updated.

Do not add the markup, run the rich-results tool, see no site-name report, and conclude the feature is missing. The documentation says that tool does not support the feature. The checks that matter are syntax, a crawlable home page, and one name that matches the visible site.

When the preferred name is not used

Google generally tries to use the preferred name from WebSite data. When it does not, the troubleshooting list is concrete. Confirm the name in the markup is the name you actually want. Confirm the markup has no syntax errors. Confirm it follows the choosing rules, including that it is not a generic keyword phrase. Confirm the rest of the home page uses that same name. Confirm you are not trying to name a subdirectory. Confirm redirects reach a page Googlebot can fetch. Confirm every version of the home page uses the same name. Then wait, and request a recrawl if you need to ask.

If that is done and the name is still not selected, add an alternateName. The system strongly considers it when it is not confident in the first name. If you still need a backup, add the lowercase domain as the last alternate. Only after that does Google suggest putting the lowercase domain in name itself. If the problem remains, Google points publishers to the Search Central help community, which is a place to report the case, not a form that overwrites the label.

None of these steps is a ranking project. A correct site name helps a person recognize the source. It does not make the article more relevant. A wrong site name, or a keyword stuffed into the source label, is a trust problem on every result from that host. Fix the home page. Leave the articles’ titles to describe the articles.

A site-name example

Northwind’s public name is Northwind. The home page heading says Northwind. The home page title leads with Northwind and then says what the company does. og:site_name is Northwind. The WebSite node is on https://northwind.example/ only, with name Northwind, url that same home URL, and alternateName NW for the acronym staff already use. They do not add “best standup software” as an alternate. They do not put a second WebSite block in the article template.

The guides live at /guides. That folder does not get its own site name. A German subdomain, if they launch one, gets its own home-page markup only if the German site should be named on its own. Until then, the domain name is the fallback Google already describes. www redirects to the bare host in one hop, and both responses carry the same name while any duplicate is still reachable. They check syntax in a schema validator. They do not expect the Rich Results Test to mention site names. They wait for the home page to be recrawled before they judge an article result.

How this shows up on a published site

BloGoose publishes articles onto the site you connect. Each article can carry its own headline, author, and dates. The site name is not a field on the article. It is the name of the home page those articles belong to. Set that once, on the home page, in words the company already uses. A subdirectory for the blog does not become a second publication in Google’s site-name system. If you want the whole host to read as your brand, the home page has to say that brand plainly.

Do not rename the site in schema to match this week’s keyword. The references on the web, and the words on the home page, are part of how Google chooses. A schema name that matches neither will lose to the domain, which is the honest fallback. Prefer the real name, keep it stable, and let the title of each article do the page-level work.

Questions about site names