Forum Moderators: Robert Charlton & goodroi
This time I prefer to keep calm about this update. So instead of start a thread about how evil-ish is Google, I decided to start a thread to discover the cause of all this mess.
I'll try to explain all SEO relevant characteristics of my affected site (I have other sites not affected by this update), and I hope more people do the same, so we can find a pattern and act consequently.
Morphology:
The site is three years old, and is structured in folder, and every folder is about a different theme and they are not related.
Every folder has articles (unique content), a discussion forum, a links section and in some cases, photo galleries.
Most of the articles have a thread in the forum to discuss about it. And the first user comments are displayed under the article. After the comments, there is a link to the related forum thread, so the discussion can go on without disturbing too much.
The article has a link to the thread, but the forum thread has no link back to the article.
I run Adsense ads in all the pages of the site.
Inbound links:
4 of the subwebs ( folders) of the site have an inbound link from 4 different DMOZ categories.
There are some (maybe 3 or 4) link exchanges, but from/to related sites.
Outbound links:
All the outbound links are to 'good' sites. The outbound links are usually only in the links section, and some directly from the articles.
Inner linkage:
The main page of the domain links to all the folders of the site.
All the pages in the folders have a link to the rest of pages of the same folder.
In addition, the footer of all the pages have links to the rest of the root of the other folders.
Every folder has a valid sitemap submitted to google a few months ago.
Special folders:
One of the mini-sites (folders) is a 'free photo album' application, so there are a lot of pages with the same text, but with a different picture.
Other mini-site is a directory of hotels and restaurants of a city in Spain, so again there are a lot of 'similar' pages.
Evolution in the serps:
The last two months the number of indexed pages in google has been growing after being in the supplemental hell.
The position in the serps for a open broad of searches was quite good, always in the first page for my targeted keywords and variations.
Panic actions:
I know I should have stayed away from making changes now, but.... Today I've created a robots.txt that exclude googlebot from indexing the images of the free photo album site, and the details of every hotel and restaurant from the spanish city site.
---------------
Any similarity with your affected sites?
[edited by: tedster at 8:07 pm (utc) on June 28, 2006]
You have to really look at paid links, do you get traffic from them? Are they really worth it? Are they paid links converting to sales? If not, why even pay for the links and take a chance at hurting your site.
Bontar, have you checked for duplicate content from scraper sites?
Frankly, we wouldnt mind so much if the sites positioned above were more relevent in the serps but the fact is they are not. All sorts of junk now ranks above our sites - i dont get it.
Im looking at some two keyword search terms where the sites in the top ten are not even on toppic!.
If this is going to stick not only will they have stuffed a load of webmasters but Googles own affiliate revenue has to be down in addition to the fact that the serps are now unusable for the search user
Just when you think Google cant get any worse!
Looking at some specific searches of the form:
"make model"
and comparing the new results with those on the last remaining unbroken DC, 64.233.189.104, for example, here is what I see:
The new algo seems to ignore <title> tags (or else give them significantly less weight).
As a result, most of the top ranked pages are the wrong pages. That is, they now seem to favor the long list of products rather than the product page itself. In one case, the highest ranked page was for the wrong product. The only mention of the "model" was in a "related products" link.
In many cases the #1 slot is now occupied by blatant Spam. I wish we could post examples, but in more than one of the cases I looked at, the top slot for "make model" was occupied by a PR0 Spam site, with no content other than:
make:model:model:model:model:model:model
make:model2:model2:model2:model2:model2:model2
make:model3:model3:model3:model3:model3:model3
make:model4:model4:model4:model4:model4:model4
In summary, this latest "update" looks to me like a return to the dark ages of search.
The site dropped only once before, late September 2005, it returned late in March 2006.
All is normal at data center 64.233.189.104
<our domain>/?NF=1
which was the way some <other site> linked to us, and STUPID Google cached it. Dumb Squats.
We never had any such page or ANY query-driven pages AT ALL.
And that cached page is 18 months old, and Google has not cleared it's cache of it. Idiots.
What an algorithm Google has! So clever, yet SOOOOOO stupid.
( Note: Query strings are ignored and resolved to the same pages as if you left out the query, so -- NO 404 is generated, so Google KEEPS THE DUMB THINGS AROUND.)
[edited by: tedster at 1:03 am (utc) on June 29, 2006]
If so, this seems to be a way to sabotage your competitors site: Just link to their home pages with query strings, then STOOOPID Google will cache the query string and say "Duhhh, that's duplicate content, duhhhhhh"
What a load of garbage. The more Google works on their algorithms, the more idiotic they get.
More trouble is something like a "custom 404" page that actually returns a 200 header instead of a 404. In a situation like that, a potentially infinite number of "bad" urls will return the exact same content with a "200 OK" http header -- and THAT situation can get your Google results all tangled up. But I've never seen a domain wide problem from a small number of duplicate urls, expecially when they originate off-site.
Our badly hit sites are taj-mahal<dot>net and others (linked from that site, if you're interested). All the virtual tour sites are pretty much the same, but for different locations, e.g. Taj Mahal, St Pauls Cathedral, and Kew Gardens.
They all have Flash-based tours, PLUS the full text (with pics) directly available to Google robots & visitors, etc. Typically this means around 35- 50 pages of description of various parts of that place (e.g. orchid house at Kew Gardens) etc. So, good, rich content on all the websites.
I cannot figure out why those virtual tour sites would be badly demoted, but the with Gardens Guide not touched.
Anyone?
When an authority site drops out the top 20 you just feel like giving up all together. You can spend over five years working non stop on a site to have google crush it in an instant because its changed its algo to incorporate something it doesnt like about your site.
Im trying to currently work out what extra algo factor google added to the results at these two data centres:-
64.233.189.107
72.14.235.107
That produces the current garabage im seeing accross the majority of data centres?
A number of pages hit are good solid static pages, no java, affiliate links etc, well written respectable keyword density etc, etc. Some other pages on the site contain affiliate links so im thinking perhaps google is applying some kind of factor to the site overall that reduces its effectiveness in google in competitive areas?
I would say go away for a few weeks and try not to panic. (easier said than done but this time I'm resisting the urge to fiddle)
I am starting to see more DC's showing the index first using the site:command on these DC's
72.14.207.99
72.14.207.104
72.14.207.107
And the results do look better as well
Just spoken to a colleague experienced in Java script.
Your travel site hit is using Java script to write html content to the page from an external file.
The garden site thats ranking fine is not doing this and is using standard java script.
We have a similar issue on pages of our own site that have been hit. Also, i did notice that since BD update the dam google bot is following ALL files including external files in Javascript so where you have src=/widgets/wid/etc within the script pointing to an external file its reading them anyway.
The significance is that google is fully aware of sites with affiliate links, pictures etc etc and has read the files
That being said you would think this factor alone would not be enough to tip the balance but maybe its a case of throw baby out with bathwater.
Your images could be something dodgy (they are not but you know what im saying)so google thinks lets give that one a downgrade just in case?
I dont know the answer to this and the Java issue may have nothing what so ever to do with your problem but i cant see any other reason why you like us have been hit - either that or google hasnt finished yet and your site will come back (either that or the garden one will get hit shortly as well!)
Good luck,
Yours a very fed up and p@ssed off webmaster
We have a similar issue on pages of our own site that have been hit. Also, i did notice that since BD update the dam google bot is following ALL files including external files in Javascript so where you have src=/widgets/wid/etc within the script pointing to an external file its reading them anyway
However, if I remove the Java or "hide" what the code is doing, then it still may take some time to recover.
I think I will try it with the St Pauls Cathedral site, since (a) it is our lowest-visitors site, and (b) it has completely disappeared from the Google serps.
Thanks - I'll let you all know what happens.
================
UPDATE --
The Explore St Paul's Cathedral and Kew Gardens websites are NOT using Java-inserted pics after all. When I looked at the code, the pictures are fixed.... So, your theory may be flawed since St Pauls is the worst hit of all our sites, and Kew is hit badly as well. (Complete removal)
[edited by: donelson at 11:19 am (utc) on June 29, 2006]
Back to the Sun, All the Best
Col :-)