When a URL no longer exits how long will it stay in the search index? Is there a way to quickly remove the invalid URL(404)?
Also when there is a 301 redirect, will the old URL as well as new URL show up in the index at the same time at all?
Thanks,
g1smd
12:36 pm on Aug 8, 2012 (gmt 0)
URLs returning 404 can stay in the index for months. You can use the removal tool to get rid of them but I rarely bother.
URLs that redirect can stay in for weeks or months and this is a good thing as the new URL might not rank yet.
piatkow
7:12 pm on Aug 8, 2012 (gmt 0)
This seems like the ******* obvious but I have seen webmasters get it wrong. ALL pages that are to be removed need to be deleted. I know of a site where the webmaster just zapped the home page. The orphaned pages just kept ranking in the top 5 positions until the hosting contract finally expired.
lucy24
10:52 pm on Aug 8, 2012 (gmt 0)
I know of a site where the webmaster just zapped the home page. The orphaned pages just kept ranking in the top 5 positions until the hosting contract finally expired.
Funny you should say that. Only yesterday-- really-- I was puttering around, let's say, example.ca/directory/page.asp exploring some stuff I found via google. All the pages had a link back to the main page-- which looks noticeably different and doesn't have a link to /directory/.
If you request a bare example.ca you get a page saying this site will be moved in 2009 to example.edu -- with link. Click on the link and, so help me, you get one of those transitional pages saying "This page has moved, please update your bookmarks, page will refresh in 10 seconds, blahblahblah..." ending up on a different page but with the exact same URL.
Possibly the site designer's ISS manual was missing a page.
But, yes, google has several pages of results for /directory/blahblah all dating from 2008 and linked from nowhere.
You don't even need to physically remove the page, if you think you might possibly need it again. All it takes is a simple RewriteRule or RedirectMatch containing the magic number "410". Or, oops, IIS equivalent.
