Forum Moderators: open
Kackle - can you explain the "dictionary" for me? And how I might benefit from it - Im reading your posts hard but dont see where youre coming from.
Sure. But you have to act quickly. Google will fix this one just like they fixed the hyphen.
1. Google is depreciating pages/sites that are over-optimized for certain keywords or keyword combinations. It does this by looking up search terms in a dictionary of target keywords or keyword pairs that it has compiled. This dictionary is Top Secret, because if you knew what was in the dictionary, you could avoid these words in your optimization efforts.
2. If the search term or terms hit on a dictionary entry, the search results for that user's search are flagged. This means that before the results are delivered, the order of the links, or even the inclusion of links, are adjusted so as to penalize pages that have overoptimizated for those terms. Most likely the title, headlines, links and anchor text are examined. It's possible that external anchor text pointing to that page has also been pre-collected and is available for scanning, but this is much less likely. (Besides, external links are not something within your immediate control, so don't worry about it right now.)
3. You want to find out which keywords that are relevant to your site are in Google's dictionary. Compile as many relevant keywords you can think of that searchers might use to find your site. Now take these words singly and in pairs, according to how users might search. Run two searches for each combination and compare the results.
4. If the results are strikingly different for the pre-filter and the post-filter search on a particular term or combination of terms, it means that some variation of those terms has been flagged because something was found in Google's dictionary.
5. Do lots of searches and you can come up with a list of "sensitive" words that you'll want to avoid when you re-optimize your pages.
It's a nice weekend project.
Altavista is looking great!
[edited by: James_Dale at 1:30 am (utc) on Nov. 24, 2003]
Bek
I edited the title, description, and h1 so that the keyword1 and keyword2 were not adjacent and left one occurrence in the body text and each separately once. The page was refreshed today and there was no noticeable effect.
Edouard_H (msg #528),
Your "experiment" failed to test a major factor in google's new algo which I know to be true.
This "factor" is that the file itself (or possibly the entire domain) is given a penalty for the specific search terms which google detects as "SEO" when the spam filter is run, and that exact file (or possibly the entire domain) can no longer rise in the rankings for that specific search phrase until the spam filter is run again and the penalty is "released" from that file (or possibly the entire domain name).
That file can still rise in the rankings for similar KW's, but not that EXACT Keyword which it was penalized for.
Even if the "SEO penalty" is assigned "on the fly" to the SERPS (as some people say is true), the actual data (from your web page) which is fed into this "spam filter" is NOT updated until the next spam filter is run, it's not updated during the rolling updates.
Therefore, this has the "net effect" of making the filter a "static penalty" and not a dynamic "on the fly" penalty which can be released by a simple re-spidering of the page data during the rolling update.
The spam filter database is a separate process from the "rolling update" which caches and re-ranks your new page data, the two processes are not tied together.
Your new page content won't be updated in the spam filter database until the next "run" of the spam filter. Some say the filter will be run each month at the normal update, I am hoping for that.
The only thing which I don't know for sure is whether the Keyword penalty is applied to the entire domain which hosts the offending page, or whether it's applied just to the exact file which was flagged for the overuse of a specific Keyword.
I am doing a test but need others to do it too...
Try testing a new file for that SAME TEST CONTENT (just switch filenames between two pages on your site), and see if your new file rises into the rankings for the penalized Keyword.
example:
file1.htm (penalized for "blue widgets")
file2.htm (penalized for "green widgets")
So just reduce the KW density of each page so that you are not likely to trip the spam filter, and then switch the filenames of both files (so that file1.htm now has the contents of file2.htm, and vice versa).
*If this doesn't get you out of the google doghouse, then that means the KW penalty is being assigned to the entire domain and you simply need to insert the contents of those pages into a different domain with decent PR (just for a test of the algo).
This test on a new domain name should give you a clean slate to form an accurate conclusion of the new algo, since the new domain will not have the keyword penalty which was assigned to the old domain.
Trying these two tests will give you some "base data" to form more accurate judgements about the new algo. I am awaiting re-spidering since I already conducted this test earlier today, but we need several people to do this test to verify the results.
I have to say that I don't think inbound links anchor text is much of a factor either - sites linking often have different-keyword1 keyword2.
You are correct that anchor text from inbound links is NOT a factor since I have seen several high ranking sites (for VERY competitive keywords) maintain their identical #1 position after the Florida Masssacre, and their only ranking criteria are 1,000+ inbound links with the same anchor text for the money word (they have less than a .04% KW density for the money phrase itself). *They are ranking because of anchor text only, and nothing else, and it's a word which is in the top 100 words being searched for on the net.
By the way, the H1 tag is also not being penalized since some of these high ranking sites use the H1 tag on every page, and thes pages all survived the Florida Massacre with #1 rankings for various popular terms.
[edited by: Brenda_J at 3:09 am (utc) on Nov. 24, 2003]
It was also in the top three from in a web position that I ran 21 Nov on google.com. (no lectures please)
That rubbish being at number three and the real estate site jumping around in the top 6 is absolutely and unequivocally not what we have had solidly for the last year.
I have no reason why you defend google so vigorously but you clearly haven't been looking at the serps for jewelry over the last year as I have been doing very intensively. I am number one for some very big jewelry related terms and a very well known player at seo in the jewelry industry. If anyone needs proof of some extremly heavyweight google number ones and top tens in this industry I have created I can provide it privately. I have optimized sites from zero to 1000 visitors per day plus for "keyword jewelry" over the last year, I didn't do that without looking at the serps on google.
The significant presence of that subdomain site in the results is a new thing, not an improvement and those results and the real estate site have not been in the top ten on google for ages. Please believe me , I really do look at the serps for jewelry every day. I am also not a newcomer as was mischievously suggested, I just don't spend much time here, I spend it working. I have no idea why nancyb said this stuff has been there a long time and then back tracked a bit and said maybe I haven't been looking lately but it just really ain't so.
To brenda - for my major kw, 3 of the top 6 sites have the kw in the domain name and/or file name.
All of my pages have the kw in the filename - I am just not showingup for one or two keywords, only more.
One thing I have noticed is that some of the links between my pages seem to be missing as far as G is concerned though that does not seem to be an influence as some with links still don't show.
The update is enough to worry about without worrying about comments being questioned. Sorry I thought it might be useful, but won't bother again in an "update" thread where everyone gets so nervouse and upset.
I think what we're seeing here is backlink importance beginning to mean a lot more.
I disagree. I have far more backlinks than 3 out of the 4 surviving competitors for my 3 keyword search terms. Most of my backlinks are even using the brandname.com domain as the anchor text. The 4 surviving competitors simply have no H1 tags. They are also using a looser variation of the Title tag than others have used to rank highly there.
The other search engines have us ranked in the top 5, but Google wants to be difficult. It just kills me because if you search for the 3 specific keywords I am referring to, then we are definately relevent. I will also point out that the top spot in overture for our keyword phrase is $5.00 a click and comparable in Adwords.
Even though we de-optimized for keyword1 keyword2 I would still think that we would have some ranking in the "new google" by having keyword1 keyword3 keyword2 etc... CACHE IS showing the new/changed/de-optimized results...
(Brenda_J -- I just read your post after writing the stuff below, and we're basically very close to agreeing. Whether the keyword precomputation is domain-specific or page-specific is a good question. You're probably right that it might be domain-specific, because that would suggest less overhead for Google and more frustration for SEOs trying to de-optimize. Seems sort of evil, though. It would also be a departure from the way Google has operated historically. They've always been page-specific in their algorithms. But then again, this new filter was a special project.)
It's possible that the process of scanning a page for the keywords it is optimized on, is something that Google has precomputed. All Google would have to do is crawl their own database and parse each English page from dot-com and dot-co.uk domains. Or even concentrate only on root directory pages, since deep pages get fewer hits. Perhaps even ignore pages where the PageRank is so low that they're not serious competitors to start with.
Extract the five or so keywords that seem most important. Then store this info under the docID. (Every page on the web has a unique docID.)
Advantages to Google:
1) You don't have to re-parse the page for this information every time it shows up in the SERPs. Since these pages are the very ones that normally rank well, they show up quite often. That means a precomputation would save significant overhead for Google.
2) You could consider anchor text in external backlinks at the same time.
3) Webmasters can't simply re-optimize and wait for tomorrow's freshbot. They have to wait until the page is reprocessed to extract the new optimized words.
Disadvantages to Google:
This will encourage penalized sites to change filenames, or even spawn new domains, so that their re-optimized pages can get new docIDs and escape the penalty.
[edited by: Kackle at 2:13 am (utc) on Nov. 24, 2003]
The key point is that it was there "before several of the latest infamous updates, I even sent a spam report to G about it after seeing it for several months - only one I ever sent." Which means that whatever it is about this site ranking well has nothing to do with a Florida algo tweak. This has to do with some older aspect of the algo.
For a dual-keyword place name which is a commercial travel destination (it has advertised hotels) adding -fdfdfd moved my page down in SERPS by about 50%. Which I assume means some commercial sites had been filtered out in the straight search.
For a dual-keyword place name which is unlikely ever to be a commercial travel destination the SERPS was unchanged.(I chose a place which was on everybody's lips during the Iraq conflict, but no one would want to visit. During the conflict the page was number 3 or so, but now it is lost among all the old news reports, etc.)
I also tested a single-keyword place name which is a commercial travel destination. With -fdfdfd my page moved down 1 position. But as it was already buried on the 7th page that is pretty inconclusive.
Looks like it's going to be tough on anybody trying to sell air tickets or hotel rooms.
Brenda, I think you are correct. I begin following this route Saturday with some of my free sites and they begin to advance with penalized keywords quite a bit. Free sites don’t have as much credibility as my regular domains though. I believe the whole domain (60% sure) is penalized because I have orphan pages in regular domains that maintain their rank. The free sites will come up high in one and two word penalized searches but the regular domains don’t.
Some of the problems I see though with changing too much are characteristic of Inktomi after Mid October. Similar to Inktomi I lectured myself not to touch anything and my site maintained its rank in Google till Wednesday. Then I had to must alter a page. Poof, domain seem to incur a heavy penalty.
There are a ton of caveats though in making this workable. I feel it will take about a week or two.
As a side note it seems to me if Google is indexing its own directory, as well as the duplicate of DMOZ, some sites could be bounced as duplicates. Since they are basically the same. What I’m wondering is how much DMOZ is affecting things.
The spam filter database is a separate process from the "rolling update" which caches and re-ranks your new page data, the two processes are not tied together.
I totally agree jmho however.
Anyway...
We got a sale from AltaVista today! I didn't even know we were ranked #1 for our prime keyphrase until now.
Google seems to have forgotten that a large proportion of searches (40% by some accounts) are commercial in nature - people looking to buy something. It DOES happen.
We don't even appear under our company name any more - a unique two word business name that can only mean they were looking for us. eg. Bumnut Enterprises.
Instead you'll find an episode of Star Trek when Barry Bumnut guest starred.
(Note: hypothetical)
Alex
I doubt that statistic is anywhere close to accurate. And, for commercial searches, from Google's perspective they would prefer you use the Adwords. ;)
SERPs look fine to me.
>By the way, the H1 tag is also not being penalized since some of these high ranking sites use the H1 tag on every page, and thes pages all survived the Florida Massacre with #1 rankings for various popular terms.
I disagree to a point. I see very few top ranked commercial sites using either an H1 or H2 tag. And when they do, they don't contain any of the search terms.
GuinnessGuy
Is majority of people here in agreement at least on the idea that current Google's results are so bad that Google is whether not done with whatever update they are doing or will have to roll back or improve it very VERY soon?
Yes, couldn't agree more! Not just for my search terms but in general for most every search I have done.
Just checked Alta Vista - #1 for all major keywords...
AltaVista DOES rock!
We're STILL #1 AND #3 on that SE.
And that goofy bras site doesn't show up at all.
Screw google.
I'm taking their search engine OFF of our site!
You and I were both thinking along the same lines at the same time, and I totally agree with your analysis.
We are starting to get some definitive traits of this new algo, we are making some good progress after just a few days of testing.
Let's keep up the testing and post our analysis as we find it.
