| Dimension | Subdomain | Subfolder |
|---|---|---|
| Authority inheritance | Generally evaluated as part of the overall site/brand today, but this is accumulated engineer commentary, not a fixed written policy | Unambiguously shares the host-level signal pool, no interpretation needed |
| Crawl budget | Allocated independently per host, a subdomain can be crawled more or less aggressively than the root domain | Shares the same crawl-budget profile as the rest of the root host |
| Topical association | Weaker by default, Google has to learn the relationship between the subdomain and the main site over time | Tighter and immediate, same host means the content is read in context of the whole site from day one |
The short answer: for most sites, a subfolder is the simpler and lower-risk choice, because it removes an entire category of ambiguity from how Google groups your content. Subdomains are not penalized and are not automatically treated as separate sites, but they carry a structural cost, Google has to do more inferential work to associate them with the root domain, and crawling is handled as if the subdomain were its own host. Whether that costs you anything in practice depends entirely on why you’re using the subdomain in the first place.
What “authority inheritance” actually means here
There’s no literal transfer mechanism where a percentage of PageRank or “domain authority” flows from root domain to subdomain or vice versa. That framing, common in SEO folklore, describes something that isn’t how Google’s systems work internally. What actually happens is closer to entity resolution: Google’s systems try to figure out whether a subdomain and its root domain represent the same overall entity, the same brand, the same organization, publishing related content under one umbrella.
John Mueller has said in multiple Search Central office-hours sessions, over several years, that Google generally tries to evaluate a website as a whole, including its subdomains, when that makes sense for the type of site. This is a meaningful shift from the older, more mechanical framing that circulated in SEO discussions years ago, where subdomains were described as “basically separate sites” for ranking purposes. That older framing was never fully accurate even at the time, but it’s especially outdated now. It’s important to be precise about the sourcing here: this is repeated public commentary from a Google engineer speaking informally about how systems tend to behave, not a single canonical written policy document you can cite with a permalink. Treat it as strong, consistent signal about intent and general behavior, not as an engineering specification.
The practical upshot is that a subdomain hosting content clearly associated with the same brand, same topic area, same overall purpose as the root domain will generally be folded into Google’s understanding of that overall site. A subdomain hosting something structurally and topically disconnected, a separate product, a separate audience, content in a different language with no crosslinking, is more likely to be evaluated more independently, simply because there’s less signal tying it to the root site’s identity.
Why crawl budget is a different question entirely
This is where the subdomain/subfolder distinction becomes concrete and verifiable rather than interpretive. Crawl budget, in Google’s own terminology covering crawl rate limit and crawl demand, is calculated and applied per host. A host, in this context, is the combination of protocol, subdomain, and domain, meaning blog.example.com and www.example.com are different hosts even though they share a root domain and even though a browser user would casually think of them as “the same site.”
This matters because it means a subdomain does not automatically benefit from the crawl budget health of the root domain. If your root domain serves fast, has clean server response codes, and gets crawled aggressively, that doesn’t automatically extend to a subdomain sitting on different infrastructure, especially if that infrastructure is genuinely separate, a different CMS, a different hosting environment, a different CDN configuration. Google’s crawlers assess each host’s rate limit based on how that host responds: response time, error rate, server load signals. A subdomain on slow infrastructure can get throttled independently of how well the root domain performs, and a subdomain on fast infrastructure isn’t guaranteed extra crawl budget just because the root domain is popular.
This per-host scoping is also reflected in how robots.txt works: a robots.txt file only governs the host it’s served from. example.com/robots.txt has no authority over blog.example.com, which needs its own robots.txt if you want to control crawling there. That’s a hard technical fact, not an inference, and it’s one of the clearest pieces of evidence that Google’s crawling infrastructure treats subdomains as distinct hosts even when its ranking systems try to understand them as part of one overall entity.
A subfolder has none of this ambiguity. example.com/blog/ sits on the exact same host as every other page on example.com. There’s one robots.txt, one crawl-budget profile, one set of server response characteristics governing the whole thing. If your root domain is crawled well, your subfolder inherits that crawling behavior automatically, because there’s no separate host boundary to cross.
Topical association and why subfolders bind tighter
Beyond authority and crawl mechanics, there’s a third factor worth separating out: how easily Google’s systems associate the content with the rest of the site’s topical footprint. A subfolder is trivially associated, it’s the same URL path structure, same host, typically the same navigation, same internal linking patterns, same templates. There’s no inferential step required.
A subdomain requires Google to build that association, usually through a combination of consistent branding, cross-linking between the subdomain and root domain, similar or complementary topical coverage, and consistent structured data or organizational signals (shared Organization schema, for instance). This can absolutely happen, and often does for well-integrated subdomains like documentation sites or blogs on established brands. But it’s an association that gets built over time and reinforced by real signals, rather than being structurally guaranteed the way it is with a subfolder.
Recommendation
If you’re deciding between a subdomain and a subfolder for something like a blog, a knowledge base, or a resource section that you want to benefit from and contribute to your main site’s topical relevance, the subfolder is the simpler and safer default. It removes crawl-budget fragmentation as a variable entirely, and it removes the topical-association question because there’s nothing to associate, it’s already part of the same host.
A subdomain still makes sense when there’s a genuine technical reason to separate it: a different application stack that can’t easily live in a subfolder given your CMS or infrastructure constraints, a separate platform with its own login/session requirements, a genuinely distinct product with its own release cadence and engineering team, or regulatory/organizational reasons to keep systems physically separate. In those cases, accept that you’re taking on an independent crawl-budget profile and a topical-association process that has to be actively supported (consistent branding, sensible cross-linking, shared structured data) rather than assumed. Don’t choose a subdomain purely out of technical convenience if a reverse proxy or path-based routing to a subfolder is achievable, since the subfolder removes structural risk that the subdomain leaves you managing indefinitely.
