What is the difference between a sub-domain and a sub-directory? In the electronic eyes of GoogleBot, not much though one is easier for the bot to deal with and the other is (in certain circumstances), easier for a webmaster to deal with.
A sub-domain prefixes the domain-name at the host level as such: ppcassurance.metamend.com/. A sub-directory is a set of files that are seen as a suffix to the domain-name such as, metamend.com/blog/.
Up until a couple of months ago, a general piece of SEO advice was to separate distinct topics addressed under the same URL to sub-domains. This was done to help search-spiders differentiate between topics and also so search-spiders would treat the information at a sub-domain as they might a unique website. Using sub-domains, SEOs could conceivably capture a far larger number of Top10 placements under the same keyword phrases for different parts of the same company. In that way, SEOs could offer five or six (or more) of ten first page placements as opposed to the more frequent two front page placements under the same phrase.
Last week at WebmasterWorldâ€™s PubCon show in Las Vegas, Googleâ€™s Matt Cutts explained to WepProNews some changes Google has made to its algorithm to decrease the number of listings from the same domain being displayed on the first page in Google results. In most cases, Google no longer differentiates between sub-domains and sub-directories though it continues recognize differences in topic between main-domain and sub-domains in certain situations.
In a post-PubCon blog post, Matt notes that it is far easier for Googlebot to detect and spider files contained in a sub-directory. It is also far easier for webmasters to set up a site using the traditional sub-directory structure. This is what Googlebot was designed to do and, with far fewer assumptions to make about topical relevance between URLs, the sub-directory structure is simplest for search engines.
For highly adventurous webmasters however, using sub-domains conveyed a possibly unfair advantage when chasing a larger share of the first page of search results. A cool SEO trick was to assign sub-domains to different host servers spread around the country or world. That way, Google would not detect and associated the same IP numbers from the host-server. The idea was to make Googlebot think it was looking at a unique website while still associating strong placements gained at the domain and sub-domain with each other. For almost three years, effective use of sub-domains was a well used â€œsecretâ€ tactic.
According to Mattâ€™s post,
â€œFor several years Google has used something called â€œhost crowding,â€ which means that Google will show up to two results from each hostname/subdomain of a domain name. That approach works very well to show 1-2 results from a subdomain, but we did hear complaints that for some types of searches (e.g. esoteric or long-tail searches), Google could return a search page with lots of results all from one domain. In the last few weeks we changed our algorithms to make that less likely to happen.â€
Interestingly, Matt reports the change in the algorithm occurred at least a few weeks ago. There has been little talk of the alteration in SEO discussion forums or social networks, indicating that most SEOs havenâ€™t noticed. The topic was addressed because someone â€œâ€¦asked his advice on sub-domains vs. sub-directories.â€
Matt Cuttsâ€™ post on sub-domains vs. sub-directories is well worth the read as are the 96 (or so) comments that follow the post.
Some made by extremely well known SEOs ask questions about sub-domains and unique circumstances such as languages (ie: fr.example.com or de.example.com).
Kalena Jordan asked, â€œWhat about sites that target different regional markets? Does it still make sense for them to use sub-domains e.g. australia.site.com, uk.site.com and use the new Regional Association tool in Webmaster Tools to indicate which searcher region each sub-domain should be associated with?â€
To which Matt replied. â€œKalena, using subdomains for stuff like fr.example.com or de.example.com is still a great approach, because those sites may be similar in idea, but the language is usually completely different.â€
Mattâ€™s answer gives a great clue into what Google wants to see at sub-domains vs. what they would prefer to see in a sub-directory. There are cases where the size of a site dictates the approach a webmaster would wish to take. Large networks that create unique content for different cities might be better served using sub-domains (vancouver.example.com/, seattle.example.com/, nyc.example.com), if information varies from city to city while sticking to a similar topical format. (ie: tourist accommodation, local search, blue widget factories, etcâ€¦)
The artful tactic of sub-domaining has likely lost a bit of the power it used to hold however many SEOs say they will continue the practice as it suits their research and deployment techniques.
In our own unending pursuit of quasi-scientific SEO research, we are taking a harder look at how we use sub-domains with an eye to improving SEO outcomes for ourselves and our clients. The traditional 4-week lull around the end of the year is a perfect time to post and evaluate experimental site structures, just to see what happens. If we see interesting results, weâ€™ll post them to this space next year.