Subdomain vs subfolder: what Ahrefs found in Mention.com

The common advice is that a subfolder beats a subdomain. Ahrefs tested that against a real migration and found the picture less clear. This write-up covers what that analysis did and did not show, and what Google’s own documentation says to do if you move a blog anyway.

What did Ahrefs actually test?#

Patrick Stox, a technical SEO at Ahrefs, looked at Mention.com, which moved from a subdomain to a subdirectory in January 2016. Using the Top Pages report in Site Explorer, he checked whether traffic rose after the move. It did not: traffic shifted to the new location, and growth began only once the company started adding additional content.

He set that against GitHub’s blog, which went the other way and then further: from github.com/blog to blog.github.com, and then to a separate domain, github.blog. He saw some flux during each transition but no evidence that any one structure was better.

What did Google say about subdomains and subfolders?#

The Ahrefs article quotes two Google voices. Matt Cutts (2012) called the two roughly equivalent and suggested choosing whichever is easier to configure. John Mueller (2017) said Google web search is fine with either, and advised picking a setup you can keep, since URL changes take time to settle.

Does Google treat a subdomain as a separate site?#

Not automatically, according to Stox. A subdomain that looks like an integral part of the site is treated as part of the main website, and he points to large brands whose subdomains appear as sitelinks under the main result. The deciding factor he names is internal linking: a subdomain with few links to and from the main site may be treated separately, and the same is true of a poorly linked subdirectory. He also notes that subdomains once produced extra listings in results until Google clustered results by host, which closed that loophole.

Why do other case studies disagree?#

Stox argues that published subfolder wins usually coincide with something else changing at the same time. He lists six confounders:

• Temporary signals. Moved pages can carry temporary site-level signals while Google works out where they should rank.
• Measurement problems. Analytics misconfiguration, seasonality or a core update landing in the same window can distort a before and after comparison.
• Blocked or noindexed pages. A misconfigured subdomain can stop signals from consolidating in the first place.
• Redesign or platform change. A new CMS, faster pages or different HTML ship alongside the URL change.
• Internal linking. Moving from a thinly linked subdomain into a heavily linked subdirectory adds links, not just a new path.
• Content changes. Pages removed, merged or rewritten during the move affect rankings on their own.

Any one of these can produce a traffic change that later gets credited to the folder structure. His conclusion is that changes introduce risk, and that switching to a subdirectory purely for SEO deserves a second thought.

What should you do if you migrate anyway?#

Treat the case as a caution, not a verdict. It has no control group, so it cannot prove subdomains equal subfolders. If you move for other reasons, such as one CMS or one analytics property, follow Google’s site move documentation, which covers subdomain and path changes:

• Map every old URL to its new URL before launch; the redirect map builder exports that list as server rules.
• Use permanent server-side redirects such as 301 or 308. Google states these do not cause a loss in PageRank.
• Update internal links to point at the new URLs rather than relying on redirects.
• Keep the redirects for as long as possible, generally at least one year, so signals can transfer.
• Expect a medium-sized site to take a few weeks or more before Google shows the new URLs in place of the old ones, and larger sites longer.

Then isolate the URL change from content and linking changes, or you will not know what caused the result. For how long full domain moves take to recover, see the domain migration recovery study.