Forum Moderators: Robert Charlton & goodroi
While investigating the cause I first checked my Google sitemap statistics particularly "Top search queries" It still reports that the site averages within the first 10 results for a handful of search terms. However, when I search on these terms rather than finding the site on the first page of results - something I have been able to do for a long time, several years at least - it is now nowhere to be found.
So I have two questions for the Webmaster World community:
Did anyone else notice a big change in search results beginning this weekend?
Can anyone suggest or refer me to methods of figuring out why my site is no longer listed in the results for search terms it traditionally ranked well with?
So - after years of methodical white hat work, doing no evil, etc the site has been relegated to the end of the results. I guess I can hope that people will often click "Next" 99 times and then visit the site.
This is almost certainly an algorithm change - possibly one based on some kind of anti-spam theory - and I would say to whoever is in the driver's seat that it looks like it was not well tested prior to going live. If that's true then I would interpret that as a sign of desperation.
Google, did you guys really just push a new algo out there and then stand back and wait to see what would happen? Ouch.
By the way for these keywords the site is still #1 on MSN and #5 on Y. I guess I need to start telling people those are the best search engines out there and sell my Google stock :p
Google, did you guys really just push a new algo out there and then stand back and wait to see what would happen? Ouch.
Indeed, i think that's exactly what happened.
I'm still of the belief that higher ups were "pushing" the algo engineers to get BD out before some type of internal deadline.
(Indicated by MC's rather vocal acknowlegement to the algo team and long vacation soon after)
And like most "rush jobs" it was not tested, refined, etc.
So now we see the past few months of them trying to figure out what went wrong.
Only a few weeks ago, a non-www check revealed that Google had not indexed any pages as the non-www version.
It shouldn't take long for stuff to start going Supplemental or URL-only if Google still works the way that it did a year ago when handling duplicate content of this type.
Sincerely, I am meticulous about 404 errors - in fact my error handling pages send me an e-mail whenever a real one occurs - and I know for a fact there is not now nor has there ever been anything like this actually on the site.
They are all of the form:
http://mysite.org/rg/deeplink-name/http://subdomain.spamsite.com/name/nm0001169/
Is anyone else seeing anything similar? Does this mean another webmaster is targeting this site? Or could the data Google has for this site be corrupted?
Meanwhile, I added "Disallow: /rg/" to robots.txt so maybe they will be cleaned up the next time G makes a pass on the site(?)
While the original site which has been out there since August of 1999 has now been relegated to the last page of results there is a blog entry dated Feb 2006 with a verbatim copy of portions of the site's original content now appearing in the #1 postion on the first page of results for the given keywords.
It makes me think I should register a new domain and convert the whole site to a series of blog entries.
I think I will check and then recheck my site today (seeing as how things are a little slow) to make sure I have no orphaned pages, dead links, etc. Good way to kill some time. (She says while making stick figures out of paper clips and waiting for phone to ring) ;)
I still see a "psuedo-page" when I do a ... site:www.mydomain.com
www.mydomain.com/?NF=1
Two other of my sites have dropped from #2 and #8 to #zillion or so...
Does anyone know if most things will recover? Or is the current sad state of affairs the new "normal"?
The last record that I have of a blip like this one happening was in March this year. It took 23 days to fully recover, but my stats finally returned to become slightly stronger that before the changes. 23 days from the 27th June would make an estimated recovery on Wednesday or Thursday this week. If nothing then we know that we have some work on our hands eh?
All the Best
Col :-)
Using site:mysite.org it appears that the site is fully indexed (around 29,800 pages) and I can account for each major section by refining the search "-pages_generated_from_database_X" etc.
Also, a brand new set of 7 pages that were just added to the site last week were found to be indexed as well with a cache date of 7/15
I originally found the new pages in the Google index by refining & paging through the site command but I can also find them by enclosing the key words in quotes so they go through the algo as a complete phrase rather than as separate words.
Further, I found that I could bring some - but not all - of the site's other pages up to their traditional results by enclosing keywords in quotes as well.
It makes it look like whatever was done has made searching for blue widgets even more equivalent to searching for widgets blue than was previously the case.
So - pure conjecture - could it be that some change took place that further decreased reliance on the keyword sequence (nearness, word distance, whatever) and consequently the overall search results are in a different order?
Literally tens of thousands of websites were hit June 27
Can you give a source for that information? (I'm not questioning the statement's validity; I'm just curious.)
I smell handcoding.I think that happens more than G wants to admit, esp. with PR.
Further test results:
Another three word phrase that used to rank around position #5 on the first page of results is now buried so deep in the results it may as well not even be there.
Also parts of the text of this page (in some cases with a link back to the original) have been copied verbatim into some newsgroup posts and some forum posts and those copies appear in the results way before the original. Google, that hurts.
One bright spot: now when I add a 4th word to the phrase from the page in question - like a refinement to clarify the context - then it is back on page one of the results where it was before.