What you explained is mostly sandbox behavior. You might rank OK for a few week when you initially launch a site but once Google's filters run, they pretty much put you into the sandbox for a period of time or indefinitely if you don't get good authoratitive links to your site.
The reason your new page ranked a bit better after the drops is that the filters may not have done their magic on the new page.
I'd be anxious to see what else you find out, I'm in a situation where it sure seems to me google has decided it doesn't like a directory anymore. Lost the ranking on the 27th also, but have been around for years. I still get some decent ranking on directory2 though ... but directory1 is nearly extinct now.
Yeah, Google is and has been degrading directories off of domains,
"My local car rental site in bla bla country" at the bottom of my site
I do have a directory for you covering all the best search topics.
You don't need to visit the directory cause it's hard to find anyways
on my site. Check all of my "link partners" sites in my directory ;)
Nobody ever does. Googles trying to rid these sites and taking down
many sites around.
What's your test showing?
It will be a couple days before the new pages get indexed. I don't think it is a sandbox issue because it is a subdomain off my pr7 site.
Would google be concerned with the depth of the directory also? Meaning
/dir1/dir2/page.html ranks better then
I'm starting to think there is something in this directory issue. My main competition has a very similar directory structure to me and was also killed on the 27th of June.
Just realized something else. Of the two articles I put up on the same day, the one in /review/ is not showing up as the first result for a very spefic sentence unless I use "". None of the reviews in /review/ are showing up for specific sentences without "".
The one that is just on the root level with .htm is coming up first for it's very specific sentence with and without "".
Still waiting on Google to index the second review in the root dir.
|Would google be concerned with the depth of the directory also? |
Not explicitly, but the new googlebot crawling patterns seem to be attentive to the PR (or click depth) of a uri. If the pages in a deep directory don't get some direct links, or at least a click trail only one or two clicks long, then they may fall by the wayside -- at least this is what several people are observing.
However, the number of directories in the url, on its own, does not seem to be a factor.
I've noticed that a lot of the pages that don't rank anymore have fallen into the omitted results. What the heck are those anyway? I went from #3 to being omitted on both "key words" and key words.
>> I've noticed that a lot of the pages that don't rank anymore have fallen into the omitted results. What the heck are those anyway? I went from #3 to being omitted on both "key words" and key words.
search for a sentence or two from an omited page. It could be a dupe issue
With "" and a long sentence search, I am 1 of 1. Without "" I am 700
Honest Reviews of Movies, Retailers, Products, and Websites
Read the review to find out more about the quality of past iPod and a look at ... Does the Rado Anatom look as good in real life as it does in the picture? ...
Rado Anatom - An Honest Review
Review of Rado Anatom. Score: 94. Sub-Section: Miscellaneous. Date: Jul 3, 2006. The Rado Anatom, while still being sold, is an older model. ...
reviews.site.com/review/Rado_Anatom - 7k - Cached - Similar pages
for all searches related to the keywords unless I include a word that is only on the /Rado_Anatom page and not on the main page. The /Rado_Anatom page has a full article about the keyword, and the main page only mentions it in a link.
This is happening with EVERY page in /review/
I can draw no conclusion other than Google has penalised this directory some how. It was doing great before July 27th and now I get 0 traffic to the pages in that directory.
The newly created page at review.site.com/page.htm is ranking normal. It uses the EXACT same format as all the other reviews.
I think this is pretty interesting.
[edited by: Soze at 12:33 am (utc) on Aug. 9, 2006]
Ok, the new page that is in reviews.site.com/page.htm is listing fine. Looks to me like there is in fact a problem with that directory.
Soze, if the site only went live the beginning of July it is much, much too early to be thinking about penalties and problems with directories. The behaviour so far sounds very normal for a new site, it just isn't going to rank number 1 in a month (or very rarely).
If your on-page optimization is good, if your navigation structure is good and spiderable, then the best thing for you to do is take Tedster's advice from above and try to build some links to the deeper pages. This will bring Gbot around a bit more frequently and give a bit of a boost to the pages' PR, which should show in a climb in rank.
There are no "deep pages" all of those are linked from the front page or one click deep. Site is just over 130 pages and it gets just under 100 googlebot hits per day. Pages always indexed within 2 days roughly.
[edited by: Soze at 8:07 pm (utc) on Aug. 10, 2006]
You say the new page is listed fine ... does that mean it's showing for your pre 27th keywords also?
|There are no "deep pages" all of those are linked from the front page or one click deep. |
Okay, I meant the ones one level down at /review/review-name. Best bet is to get some links to them or a heck of a lot of links to your home page.
As Lothar said way at the top:
|What you explained is mostly sandbox behavior. |
And this is mostly normal for a new site. Things are going to bounce around for awhile.
By listing fine I mean that it is showing for the keywords that are associated with it. These arent the same keywords as the dropped pages. I will do a test on that tonight.
Thing is, new pages about new keywords in /review/review-page are the ones having problems. New or old, they all get no listings. So far I am 2 for 2 in getting decent spots for these new pages. The idea that the keywords are different and may be easier or harder to rank for is unlikely. They feature the same, if not more competition than the others.
Also, why would google sandbox a subdomain on a pr7?
|Also, why would google sandbox a subdomain on a pr7? |
Because a new subdomain is pretty much seen as a new domains -- no history, no established trust or quality.
Ok, I just did a test with very interesting results. I took a page that was in the "penalized" directory, copied the source, made it a static .html page and placed it in a different directory.
It is now listing normal in Google. It was 99% the same.
I'm doing more experiments tonight.
<edit reason: Forum Charter [webmasterworld.com]>
[edited by: tedster at 1:15 am (utc) on Aug. 24, 2006]
While there might be some logic in G going after entire sections (I can think of numerous reasons for it that would make sense), I don't believe it's happening.
What I have seen, is that often a whole section is in trouble because the pages in that section are too similar to one and other, or the link structure is botched up, or someone forgot to individualize the title tags in that section, etc.
Are you sure that your scripts are working properly? You said you copied the code and made static HTML and it started ranking well. Maybe the script does something wrong, and this is the reason, not the directory - except the fact that content in this directory is generated dynamically?
Use Live HTTP Headers or similiar tool to check if HTTP Headers for dynamically generated files are made properly - for example, some CMS systems add redirects which would kill Google ranking.
And adding links from other domains to your deep pages (saying deep pages, I mean pages other than home page) is essential.
I've had A LOT of my pages completely drop from the google. Searching for URLs doesnt return anything.
But the sites index.html page still rank #1 for some very competetive keywords.
i think the directory /review/ is pointed to the keyword 'review'
if a directory contains more than 1 review the directory must be logical called /reviews/
if your site has no spam, or other black hat things.. just wait to get out of the sandbox.
The sandbox is located in the sahara and its Google his or her problem. Not yours.
I can definately say that the pages are unique and so are all the title tags. Of course there are template consistancies, such as score, date, and company names, but the text on these page is completely different.
Link sturcture is fine. You can hit these pages many different ways.
I don't see how the scrupt could be causing the problems because the pages are still indexed and pre July 27th, rank was awsome. I don't see the script breaking over night.
The idea that /review/ should actually be /reviews/ makes no sense. I see no evidence that Google would read my directory as only having one page ad then penalize it for having more.
Consider Rotten Tomatoes:
site.com/m/snakes_on_a_plane/ That would be like saying, m should have an s at the end.
The only other thing I can think of is that the is a problem with not having a trailing /.
Either way, I'll go forward with more tetsing and see what happens
Heard that before, so will ask the question: Do you mean that the reviews were written for your site only and are completely unique on the Web?
Also, anything in how the review pages are generated/structured? For example, review sites are a dime a dozen these days; being cranked out like crazy the same way directories were two years ago before they started getting killed in the SERP's. Perhaps a too common page profile relative to others on the Web (especially if generated by some kind of software).
IMHO, G tightened up dup restrictions recently on a number of fronts ranging from page structure to titles to (possibly) internal nav.
All reviews are written by me or my friend. All 100% exclusive to my site. Only consistancies in page content is layout. For the most part though, nearly all of the text is different. Only navagation links are similar and words like score: Company: Date:
Many sites maintain layout. It is in the nature of CMS to maintain templates that dynamically display different data within the template.
|The only other thing I can think of is that the is a problem with not having a trailing /. |
And does your server respond the same page if tried an URL with trailing /? It would be a duplicate. Myself, I use paths like /directory/article_title.html for dynamic content - I find .html extension safer.
As for layout similiarities - it depends what percentage of page code is layout, comparing to the content. If layout, and, what's worse, navigation, takes place in the beginning of HTML code, it sometimes is considered duplicate content by Google. So to be safe, it's better to make minimum layout in the beginning of the template, and start the content as soon as possible. The ideal code would be like:
You can use CSS to adjust its appearance, and place some headers/navigation to top, but in HTML code it's better to begin with most important content to make it easier for SE to decide what's the subject of the page. That's my opinion and it works well for me, correct me if I'm wrong.
|If layout, and, what's worse, navigation, takes place in the beginning of HTML code, it sometimes is considered duplicate content by Google. So to be safe, it's better to make minimum layout in the beginning of the template, and start the content as soon as possible. |
Hmm…interesting, but why? SE’s processing system can strip all HTML and look at pure text content. Granted, SE’s algos can look at layout properties such as bold, underline, etc, and incorporate it into evaluation mechanisms, but this is separate (or in addition to) looking at the text, so to me physical positioning of text and layout wouldn’t matter much. Is there any empirical evidence of this? [not trying to be dismissive, just curious]