Forum Moderators: Robert Charlton & goodroi
I have my opinion - what's yours?
I would love to hear what morsels of info led you to the conclusion that you have come to.
In my brain I can't see an intentional line of code that would consider the length and give it a +1 or -1 based on that length. However I can see them using a string buffer that complains about long URLs and as an unintentional result causes it to not get a +1 rating for "good URL" or something like that...... all the above is speculation.
Does url length still affect rankings?
No. But, URI structure does. Proper structure in turn reduces the total number of characters so they are by default, shorter.
URI length affects usability, that's for sure.
Dymanic i don't touch those....
Vimes.
Does url length still affect rankings?
Not sure exactly what is meant by "url length."
If we're talking about the number of directory levels from the root, I've never thought that mattered.
What does matter (still matters) is the number of clicks from the root... ie, the link hierarchy... which may or may not parallel the directory structure. When I say "the root," btw, deep links can shift the top of the hierarchy... but generally most inbounds come to your default page.
You can short circuit this hierarchy, as Wikipedia and DMOZ do, with linking that jumps across the strict hierarchical structure. On a large enough site, you may even end up with a sideways structure rather than one that's top down.
If you have a blog on blogspot.
Based on discussions I've read, blogspot sites get preferential treatment out of the gate. ;)
Length does not affect rankings. What the URI contains may be the determining factor, not the actual length. I have examples of both that are doing just fine.
Please, define length in this instance. What is too long? If there is a link on your home page that goes 10 levels deep, it is going to rank just fine due to the click path.
Does url length still affect rankings?
Haven't noticed alot of figures yet as to what each of you consider as being "long" and what is acceptable?
To clarify, we are talking about whole URLs and not domain names, right? If so, I would say that the type of page would be a factor:
/readme.php?source=direct&page=34&articletype=widgetingforests&loc=uk
would rank better than:
/readme-source-direct-page-34-articletype-widgetingforests-loc-uk.html
or
/readme/source/direct/page-34/articletype/widgetingforests/loc-uk/
as the latter two show either the use of a bad mod_rewrite (bad because it doesn't really make it more usable) or is probably trying to spam in as many keywords as possible into the file/folder names. Obviously .php versions can be spammed too:
/readme.php?spammyscraper=insurance+services+for+people+with+red+hats+and+pink+slippers
but if a search engine can't spot that then there is something wrong with them! ;) So no, I've never noticed with my sites (I have 2 or 3 of each only so it's not the greatest test!) that length was a problem, only how that length was used.
Mike
/readme/source/direct/page-34/articletype/widgetingforests/loc-uk/
using your example, i have a theory that if none or few of those "directories" actually resolve to something meaningful, then the entire url is disregarded or not ranked very highly.
that is, if you don't have any content in...
/readme/source/direct/page-34/articletype/widgetingforests/
/readme/source/direct/page-34/articletype/
/readme/source/direct/page-34/
/readme/source/direct/
/readme/source/
and
/readme/
... then it's obviously an attempt at tricking crawlers (IMHO)
my company have lamely attempted to "just replace all ='s ?'s and &'s with slashes" (with very loud protestations from me) and have seen naught from the effort. so now we've lost ranking for the old ugly urls, and three months on haven't had the new ones indexed.
/readme/source/direct/page-34/articletype/widgetingforests/
Looks to me that this scheme is triggering off "page-34" to find the correct database row. I'd suggest stress testing those urls and find out if your server resolves something like this:
/readme/source/direct/page-34/articletype/any-old-crapola-here/
However, that url is not anywhere near long enough to be a problem just because of length. If there is a problem with long urls, I think it would take hundreds of characters to trip it.
Brett, can you define what you mean by long URLs? Are you talking about long because of many folders deep or long because people are putting too many words with _ or - in the URL?
Does url length still affect rankings?
I've got some pages with very long URLs and they rank in the top 10.
It all depends on what you mean 'affect rankings'. Like freelistfool said, long URL rank fine in top 10; however, how easy it is to get them ranked is where I would raise a question. My experience has been that though long URLs can and indeed and do rank, it is more difficult to get them to rank. It takes the engines longer to 'trust' longer URLs.
Can we start a thread, derivative from this thread's question:
How do you choose your URL file names?
p/g
Yahoo tends to truncate URLs if they are too long and when the scrapers come along to Yahoo generated SERP and use the truncated URLs and post the junk on their sites and it creates the junk on your site if spidered by SEs.
As far as length & ranking, it does not matter, we have top 10 comprised of ids=%$&%&%h and wholly root folders.
So my hint on this would be: URLs don't Rank, unique content does.
Matt Cutts' blog that he recommended it's better to have less than two "=" or "&" in the URL
Why say less than two when he could have said one, which sounds less cryptic? Are you sure it's not no more than two?
Why say less than two when he could have said one ...
I should write "not to have more than two" instead of "to have less than two". My bad.
internetheaven
/readme/source/direct/page-34/articletype/widgetingforests/loc-uk/
using your example, i have a theory that if none or few of those "directories" actually resolve to something meaningful, then the entire url is disregarded or not ranked very highly.
I agree completely, I've noticed that even if they are not linked to, Google tries to crawl the folder itself which is why I now create an index.php file for each and every folder even if it's just for show.
I have read in the Matt Cutts' blog that he recommended it's better to have less than two "=" or "&" in the URL.
Wasn't that some time ago? Has that been said again recently because I've noticed more references to advancements in dynamic URL indexing over the past year by Google.
New Theory - one thing I tested (only on one site mind you but it was quite dramatic) was the % of pages with long URLs that affected overall ranking of the entire site. I was toying with it after Google FINALLY started to remove all those search engine scraper pages from the top rankings on google.co.uk - their sites were entirely made up of long URLs because they were auto-generated spam pages, nothing else. I switched a 500,000 page site from over 90% long dynamic URLs to at least a third of that using a mod_rewrite and some PHP scripting. Rankings increased substantially across the whole site on Google (Yahoo and MSN were unaffected).
Which leads me to another question - why did you start this thread in "Google Search News" Brett? Something specific you want to share with us about a new patent Google have applied for recently?
Any search engine favoring shorter URIs by 2020 is going to be missing a load of potentially good sites. Surely new sites created after 2008 won't all be spam sites?
If short URIs are preferred by search engines then the search engines have become confused about the whole web thing. URIs are things used to locate pages (or indicate where a page might be found).
Maybe search engines will develop the technology to have a look at pages before dismissing them due to their long URIs.
However, that url is not anywhere near long enough to be a problem just because of length. If there is a problem with long urls, I think it would take hundreds of characters to trip it.
I just checked. I have a URL that is 114 characters long (including the [)...] that is ranking top 10 on a 3 word term alon the lines of 'state widget makers'
The URL (er, URI, pageone ;)) is along the lines of
www.example.com/industry-directory/widget-makers/us-widget-makers/state-widget-makers.htm
It is NOT a competitive term.
However, a quick look at 'city state web design' showed a number of very similar URLs - though perhaps not 114 characters :o
Depending on the IE version, URLs over this length either become unclickable or return a generic error.
Not something that effects many sites but it does happen sometimes when there's data (such as session or forwarding info) encoded in the URL.
I mean: I'd accept the argument that people from time to time type in domain-names in the browser bar, but full URIs? Do screen-readers for blind people really read out the whole URL?
> Google tries to crawl the folder itself which is why I now create an index.php file for each and every folder even if it's just for show.
I'd second that and did the same a few months ago. However, with no visible effect on ranking yet, but a much tydier broken-links-section in webmaster central.
