Google Search Advocate John Mueller recently described a case where a hidden HTTP version of a homepage caused Google Search to show the wrong site name and favicon. The incident, described on Bluesky, highlights technical differences between how browsers and Googlebot handle HTTP and HTTPS.
Hidden HTTP homepage affecting Google site names
In the case Mueller detailed, the site correctly used HTTPS for normal traffic, but a server-default HTTP homepage was still accessible in the background. While visitors' browsers upgraded HTTP requests to HTTPS, Googlebot continued to crawl the separate HTTP version of the homepage.
Mueller explained that this setup produced conflicting homepage signals for Google's site name system:
- Chrome automatically upgraded HTTP navigations to HTTPS, so users rarely saw the HTTP homepage.
- Googlebot did not apply the same upgrade and instead used the HTTP homepage when evaluating the site name and favicon.
- The HTTP homepage returned server-default content, not the intended HTTPS homepage content or structured data.
He wrote that this was "a hidden homepage causing site-name & favicon problems in Search" and called the situation "a weird one."
Detection methods and Googlebot behavior
Mueller advised site owners to test what Googlebot retrieves from the HTTP version of their site, using tools that bypass browser-level HTTPS upgrades. These checks reveal the raw HTTP response that Googlebot may be using for site name and favicon selection.
Specifically, he wrote that useful diagnostics include:
- Running
curl http://yourdomain.comfrom a command line to inspect the actual HTTP homepage response. - Using Google's structured data testing features in Search Console to analyze the HTTP URL.
- Using the URL Inspection tool's Live Test to see what Google retrieves and renders for the homepage.
Because Chrome and other modern browsers upgrade many HTTP requests to HTTPS automatically, these hidden HTTP pages may not appear during normal manual testing. Googlebot, however, can still access and use the HTTP version directly if it is available.
Background and official documentation
According to Google's site name documentation, its site name system relies mainly on homepage signals when choosing labels in search results. These signals include WebSite structured data, title elements, headings, and properties such as og:site_name.
The same site name documentation notes that some sites effectively have duplicate homepages, such as separate HTTP and HTTPS versions. To avoid conflicting signals, Google recommends providing matching structured data and consistent naming on each homepage version.
Google also specifies that WebSite structured data should appear on the domain-level homepage, which it treats as the primary reference for site names. Site names are not currently supported in the Rich Results Test, according to the site name documentation.
The case and its technical details are documented in Mueller's public wrote post on Bluesky and in Google's official Search guidance. Key sources include:
- John Mueller's Bluesky post describing the HTTP homepage issue and its impact on site names and favicons in Search: Bluesky.
- Google Search Central's guidance on site names, structured data, and duplicate homepages: Site names documentation.
.svg)
.svg)
.svg)