We probed 232 ranking domains on 31 July 2026: 87.9% keep the blog in a subfolder, but 66% run docs on a subdomain. See the split, then pick yours.
Across the 232 domains we sampled from live Google results on 31 July 2026, the answer splits by content type. 87.9% of blogs sit in a subfolder, and blog subdomains are folding into subfolders roughly nine to one. Docs invert it: 66% run on a subdomain. Subdomain vs subdirectory has no single winner.
Quick verdict: Put your blog, guides and resources in a subfolder — that's where the ranking web has already moved, and it's still moving. Keep docs, help centres and standalone apps on a subdomain when a vendor platform requires the CNAME. Storefronts are a genuine toss-up; decide on platform constraints, not SEO folklore.
| Dimension | Subdomain (blog.site.com) | Subdirectory (site.com/blog) |
|---|---|---|
| How Google handles it | Often crawled as its own site | Part of the root site |
| Signal consolidation | Builds separately | Pools with the root domain |
| robots.txt | Needs its own file | Root file governs it |
| llms.txt | Needs its own file | Root file governs it |
| Search Console | Separate property | Inside the root property |
| Analytics | Cross-domain setup needed | Works by default |
| Tech stack | Can be fully independent | Must share routing |
| Our scan — blogs | 8.1% of 124 domains | 87.9% of 124 domains |
| Our scan — docs | 66.0% of 53 domains | 30.2% of 53 domains |
| Our scan — shops | 48.8% of 43 domains | 46.5% of 43 domains |
| Strongest use case | Vendor-hosted docs and help | Blog, guides, product content |
Published advice here contradicts itself. Cloudflare's engineering blog argues that the subdirectory strategy results in better root domain authority; several vendor help pages insist subdirectories have no SEO advantage at all. Both sides argue from anecdote. I build SEOmator's crawler, so rather than pick one, we measured what ranking sites shipped.
In simple terms, a subdomain is a separate unit or partition of your main website, with its own unique content. It acts as a distinct branch of your main website.
For instance, if your website is www.chocolates.com, a subdomain could be whitechocolates.chocolates.com.
You may want to start a blog relevant but different from your main business content, or launch a separate e-commerce store apart from your main marketing site. Here, a subdomain like blog.chocolates.com or shop.chocolates.com would come into play.
Subdomains let you segment portions of your website according to different needs or topics. You might opt for a language-specific subdomain to target a specific geographical market, or an event-specific subdomain for a temporary marketing campaign.
Knowing when to apply subdomains will ultimately shape their impact on your SEO.
If you run a large e-commerce site with diverse products that require dedicated content, or manage a multiregional, multilingual website, separation with subdomains might serve you better. The same goes for anything on someone else's platform — a hosted help desk, a hosted docs product, a status page.
The conventional advice adds one more case: a corporate blog that wants separate branding and its own SEO effort, on blog.yourdomain.com. That's the single case our scan does not support.
The independent nature of subdomains means any SEO advantage or disadvantage stays largely on that host. A rise in blog.yourdomain.com's rankings doesn't reliably lift your main site with it.
The genuine wins are operational, not algorithmic. A subdomain hosts content distinct from the main site, runs a different stack without touching the main site's routing, and lets you experiment without exposing your money pages.
The costs are just as real. Every subdomain needs its own SEO effort, because its wins don't automatically reach the main site, and authority builds separately rather than pooling. Content duplicated across hosts also invites canonical problems, where search engines pick one URL and quietly drop the other.
A subdirectory, also known as a subfolder, is just a folder within your main website. It is used chiefly to organise the content of the website.
Subdirectories are appended to the end of the main domain. For example, www.chocolates.com/darkchocolate is a subdirectory of the main website.
Your 'About Us' or 'Product' pages typically fall into subdirectories like www.chocolates.com/about or www.chocolates.com/products.
Search engines treat content in a subdirectory as content of the primary domain. Link equity flowing into a subfolder page joins the same internal PageRank pool as the rest of the site, rather than starting a second one.
The primary consideration should be the nature of your content.
If the content across your website has a strong correlation and supplements your overarching theme, then opting for subdirectories might be the path to tread. An e-commerce site with product categories, a blog sorted by topic, a news site sorted by section — all work well in a subfolder layout.
The impact of subdirectories on your SEO relies on maintaining a clean, well-structured site. Serving high-quality, value-adding content is what makes subfolder work lift the whole domain rather than drag on it.
Subdirectories create organised sections inside one site, so your SEO effort lands on a single entity instead of being split across properties. Internal links, sitemaps, robots rules and reporting all stay in one place, and simpler URLs make internal linking obvious.
The trade-offs: a clean subfolder architecture is harder to build on very large sites, and reputation is shared. Thin pages in a subfolder drag on the same domain your money pages depend on — which is exactly the coupling that makes subfolders powerful when the content is good.
We took the 232 domains ranking on 38 live Google SERPs and resolved both candidate locations — https://<type>.<domain>/ and https://<domain>/<type> — for four content types. 928 host probes, and the redirect chain is the finding.
Blogs are not a close call. Of the 124 domains that had a blog at all, 87.9% served it from the subfolder side and 8.1% from the subdomain side. Only six in the whole sample — Adobe and Coursera among them — still run a live blog subdomain with nothing at /blog.
Docs go the other way. Of 53 domains with documentation, 66.0% were on a subdomain and 30.2% in a subfolder. Help centres lean the same way at 59.2%. Storefronts refuse to pick: 48.8% subdomain against 46.5% subfolder.
A live redirect tells you which way a company already voted. When blog.example.com answers with a 301 to example.com/blog, someone did that migration on purpose and still pays to maintain it.
37 of the domains we probed redirect their blog subdomain into a subfolder. Four push /blog out to a subdomain. Roughly nine to one, in one direction, among sites currently ranking — and the names on the nine side aren't small: Atlassian, Slack, Zapier, Datadog, New Relic, Elastic, Intuit, 1Password, Backblaze. On the four side, HubSpot, LastPass and Xero still send /blog to a host of its own.
Because most of those subdomains were never an SEO decision. We fingerprinted 72 docs and help subdomains from the sample by their response headers and HTML signatures. 40 of them — 56% — run on a third-party platform: Zendesk (11), Intercom (8), HubSpot (7), GitBook (6), Salesforce (4), Mintlify (3), ReadMe (3).
Those platforms hand you a CNAME target. You point help.yourdomain.com at their infrastructure and you're done. There's no subfolder option to evaluate, because serving someone else's app from /help means building and maintaining a reverse proxy. The subdomain isn't a choice about link equity. It's the shape the vendor ships.
So "66% of docs are on a subdomain" is not evidence that docs rank better there. It's evidence that docs get outsourced — the most common instance of this decision in SaaS technical SEO, usually settled by procurement before SEO is in the room.
The honest caveat: 32 of those 72 hosts showed no third-party signature. Oracle, Datadog, Palo Alto Networks and 1Password genuinely self-host their docs on a subdomain and could have chosen otherwise. Vendor lock-in explains most of the inversion, not all of it.
Mostly yes, for practical purposes. Google associates subdomains with their root domain and attaches no penalty to using one. But the systems you touch daily treat them as separate: crawling is scheduled per host, robots.txt applies per host, Search Console reports per host.
That's the gap between "Google says both can rank" and what happens on your site. Both can rank. One of them starts with the internal links, crawl history and reporting you already have.
This is the part nobody puts in the comparison posts, and the part I deal with most. Every hostname is a separate crawl target with its own configuration surface.
| What you now run twice | Why it matters |
|---|---|
/robots.txt | Only governs its own host |
/llms.txt | Same per-host rule |
| Search Console property | Separate coverage and query data |
| XML sitemap | Each host submits its own |
| Analytics config | Cross-domain tracking + cookies |
| TLS certificate | Own cert or a wildcard |
| CDN and cache rules | Own origin, own headers |
| Canonical and redirect policy | Two hosts to keep in sync |
The robots file is the one people get wrong most. Google's spec is explicit: "The rules listed in the robots.txt file apply only to the host, protocol, and port number where the robots.txt file is hosted," and a robots.txt on a subdomain is only valid for that subdomain. Your carefully tuned example.com/robots.txt does nothing for blog.example.com.
The same per-host rule extends to the files that govern AI crawlers. llms.txt is fetched from the root of a host, so blog.example.com/llms.txt and example.com/llms.txt are two files with two scopes. Publish one, forget the other, and half your content is described to answer engines by a file that doesn't cover it.
Crawl permission works the same way. Whatever you decide about AI crawlers like GPTBot, ClaudeBot and PerplexityBot, that decision is expressed per host — so a split architecture means making it twice and auditing it twice. The pattern that costs people AI visibility is rarely a deliberate block. It's a second host nobody configured.
Treat it as a site move, not a URL tweak. Google documents this path directly, and the mechanics matter more than the decision.
Decide per content type, not per site. Here's the shape our data supports:
| Content type | What our scan found | What I'd ship |
|---|---|---|
| Blog, guides, resources | 87.9% subfolder | Subfolder |
| Docs, all sampled | 66.0% subdomain | Depends on the next two rows |
| Docs on a vendor platform | 56% of docs/help subdomains ran third-party software | Subdomain — no real choice |
| Docs you host yourself | Not separable in our scan | Subfolder |
| Help centre | 59.2% subdomain | Follow the helpdesk vendor |
| Storefront | 48.8% vs 46.5% | Either; decide on platform |
| Language versions | Not measured | Subfolder plus hreflang |
| Standalone app or dashboard | Not measured | Subdomain |
Two need a note. For language versions, subfolders keep every market on one authority profile, and self-referencing hreflang does the targeting a separate host used to do badly. For storefronts our scan found no winner: if your commerce platform serves /shop cleanly, take the subfolder; if it can't without a proxy layer you'll regret, take the subdomain and move on.
The structure decides where signals pool. It doesn't decide whether you have any.
Publish content worth the URL and update it when it goes stale. Keep the design responsive — check any host with our free mobile-friendly test. Use target terms in content, headings and URLs without stuffing them, which our free keyword research tool makes easier to plan. Link internally with descriptive anchors, and monitor crawl coverage and rankings per host rather than trusting what the structure promised.
We scraped 38 fixed Google queries (US/English) on 31 July 2026 spanning B2B SaaS, e-commerce and publisher intent, and collected every unique registrable domain ranking organically. Platforms and aggregators — Reddit, YouTube, Wikipedia, Amazon, G2 — were excluded. For each of the 232 remaining domains we resolved both candidate hosts for blog, docs, help and shop and followed the redirect chain: 928 probes. A second pass fingerprinted 72 docs and help subdomains. The SERP layer is the same one behind our Google SERP rank checker.
Four limits, stated plainly. This is a one-off SERP-sampled scan, not the SEOmator customer panel — nothing here describes sites we monitor for clients. Presence was detected heuristically at /blog and blog.<domain>, so a site publishing at /resources counts as having no blog; that understates how many sites blog, but doesn't bias the split we report. It's a single snapshot, so we show the direction sites have migrated, not the rate. And fingerprinting is signature-based: a self-hosted docs site embedding a HubSpot script can register a false positive.
For practical purposes, yes. Google associates a subdomain with its root domain and won't penalise you for using one, but crawling, robots.txt, llms.txt and Search Console all operate per host. A new subdomain starts colder than a new folder on an established host.
A subdirectory. 87.9% of the 124 sampled domains with a blog serve it from a subfolder, and 37 sites in our scan have already redirected their blog subdomain into one against only 4 going the other way. The migration traffic is almost entirely one-directional.
Map every URL 1:1, use 301 or 308 permanent server-side redirects, keep them at least a year, rewrite your internal links to the final URLs, verify both hosts in Search Console, and submit a sitemap of the new URLs. Expect a few weeks of fluctuation on a small-to-medium site.
They aren't inherently. What they cost you is consolidation and overhead: authority builds separately, and every per-host file and property has to be configured, monitored and kept in sync twice. On a small team that duplication is where things quietly break.
blog.chocolates.com is a subdomain — a separate host. chocolates.com/blog is a subdirectory — a folder on the existing host. Same content, different addressing, different consequences for crawling and reporting.
In https://blog.example.com/seo/guide, blog is the subdomain and /seo/guide is the path. Everything before the first single slash after the protocol identifies the host; everything after it is the path on that host.
Links to a subfolder page feed the same internal PageRank pool as the rest of the site, so equity spreads across the domain through normal internal linking. Links to a subdomain build that host's profile; cross-host links still count but don't consolidate the same way.
Third-party authority metrics are usually calculated per host, so a subdomain typically scores independently of its root. You can check both with our free domain authority checker. Treat the gap as a rough signal of how much link equity you'd be starting over on.
Three defensible reasons: a vendor platform requires the CNAME, the content needs a genuinely different stack, or it's a separate application rather than content. Branding alone isn't one — that's the case our scan found sites reversing.
Both structures rank. That's true, and it's where most articles stop — which is why the question keeps getting asked.
The verdict from the data: for content you write, own and want ranking — blog, guides, resources, product pages — use a subdirectory. It's what 87.9% of sampled ranking domains do, the direction 37 of them already migrated, and the only option where robots rules, sitemaps, reporting and internal links stay in one place. For content you don't fully control — vendor-hosted docs, a hosted help centre, a separate app — use a subdomain, because the platform already decided and proxying someone else's app costs more than the SEO does.
The subdomain vs subdirectory debate isn't unresolved. It's resolved differently per content type, and ranking sites have acted on that quietly for years.
Not sure which side your structure is on? Run a free SEO audit against both hosts and compare what comes back.
🔍 Related articles:
- Website Migration: Your Ultimate Checklist and Guide
SEOmator Rank Tracker
Technical fixes are invisible until positions move. Track the keywords a fix was meant to affect and watch whether it actually landed.
SEOmator Rank Tracker
Benchmark your site against real 2026 data: Cloudflare Radar crawl ratios, Web Almanac Core Web Vitals pass rates, and post-AI-Overview CTR — every figure sourced and dated.
Only 45.9% of crawler requests return a 200. The 2026 crawl waste report breaks the waste down by client type, industry, and operator — and finds AI bots are the cleanest traffic on the web, not the dirtiest.
Discover the top 10 technical SEO issues affecting websites in 2025, with data-driven solutions. Learn how to fix crawlability problems, improve page speed, resolve duplicate content, and implement schema markup to boost your search rankings.
