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.
Anyone care to give me an update?
Is the dance over?
Keyword1 Keyword2. INFO Keyword1 Keyword2. Is it real? I love you MORE INFO keyword1 keyword2 keyword3 keyword4 ... For jesus lives in every shot If you believe
it becomes real Do you ... I'll talk to god I've acted out my love in stages ...
MORE INFO keyword1 keyword2. Please let me know that it’s real Out of the underground
Everything became so ... re not my first love So take my hand I love you in a ...
Remember: this site is #2 for one of the biggest money terms on the internet. It has remained in place throughout Florida. There is just no consistency here...
dazzlindonna,
I've even tried replacing "cheap" with "adult", as some posts suggest that an adult filter is being applied; it still gives on-topic serps. Even for "cheap adult widgeting" "buy adult widgeting" or "free adult widgeting" i find on-topic serps
Perhaps we're not talking about the same type "widgeting" - the one i referred to is a product that every webmaster needs, unless they run their own server or are an ISP.
/claus
I have tried to figure out what has gone wrong, but it appears that the results are at best random, as what ever theory you come up with using one keyword set, does not work with another keyword set.
I can only conclude that this update is so messed up, that
A) the update is far from complete and things will get much better in the next couple of days.
B) The Google algorthim has gone totally haywire, and google will try for a few more days, before throwing the whole thing out and revert to a pre-Florida index.
Up until now, I have been rewarded for not using SEO techniques, but concentrating on content. It may not be today or tomorrow, but I expect that I will be restored to my former position once this whole mess is over.
Also of note, it appears tha -va has switched over to the
garbage results. Has anybody seen a datacenter, that shows
some glimmer of hope, now that -va is gone over to the dark side?
if you do your same search with -dkidkdid after it, you'll see a difference. for my particular nichekeyword, it returns pre-florida results. not sure if yours does or not. but if so, you can probably see that both are relevant returns and then analyze both to see why they are different now. if you come up with anything, let us know.
No, but Altavista results are looking much, much better than Google
Always liked Altavista for their advanced search options. On the other hand, if they were still #1, I'm sure there'd now be lots of complaints about their updates.
By the way: haven't seen very much on optimizing for search engines other than Google for a while...
Laurenz
I can't check H1s etc, since the page is cloaked. Not sure how to access the actual cloaked page.
I quite like the SELG site, but I'm thinking about building something a little bit more informative, with more content, etc. Any takers?
What I am seeing is that the keywords have to be in the title and H tags for this to work.
I'm a Yank, not from Bristol that is my name... ;-)
All previous seo techniques are gone!
There is no dictionary!
It is not about throwing out ecommerce sites since that would mean google should get thrown out too (they do sell advertising don't they)!
It is not about over-optimizing!
It is not about any of the hundreds of other theories!
We know that Google has been interested in AI. We know the issues with dmoz, which has been in the past the reference base. We know GG has said many times before they are more interested in "automated" methods of spam reduction & relevancy scoring.
What we may be seeing is the first of AI as applied to web searches. Ok so what about the inconsistencies etc. Well we have to train the new AI brain. This would work much like a Bayesian filter works ( spam reduction using statistical probability).
If you browse through a keyword listing extensively, some general conclusions come to light. Some listings are relevant, some are close and some just do not belong there. As is common in the statistics world the phrase "95% confidence level" applies which means that 5 % simply doesn't even make sense (these are our listings that don't make sense).
So if this AI theory is right then over time results will get better as the AI brain is more fully trained.
Just trying to make sense of what has become a slippery slope.
These sites had about 300-400 of excellent content related to "city keyword". No spam technics were used. I can say that my competitors had a pretty similar set up and even-though they are my competitors I must be honest and say they also had excellent content and didn't spam.
Now that these sites are removed the first two pages of serps are dominated by not related to industry sites. Mainly sites which library, newspaper, university, or radio sites that may be mention once or link to a "city keyword1" site. This could imply that pr is becoming a dominate ranking element.
I personally believe Google is filtering optimized pages. The problem I see with doing this is sites that are optimized are generally very relevent and usually have excellent related content. SEO is not a bad thing we (the SEO and search engine) are both working for the same goals. It is a shame that Google does this so close to Chirstmas. I imagine thousands of site owners will be affected.
I'm seeing exactly the opposite.
Tha't what is so confusing about this update, every theory can be disproved by another set of keywords.
In my main keyword phrase
PR1's (I never knew that there was such a thing) are in top 20 of a very competitive category that used to be dominated by PR5 - 6's and an occassional PR4 but never below PR4. Now, in the top 20 results there are 2 PR1s, 2 PR2s, 4 PR3s, 10 PR4s, 1 PR5 and 1 PR6.
If the cause is 'over optimisation' - well, we're smart people - smart enough to de-optimise!
I think maybe it is a second update - and we're just going through the nonsense stage again.
<edit: clarification: by 'nonsense' I don't mean the contents of this thread - I mean Google throwing up nonsense! >
[edited by: superscript at 3:26 pm (utc) on Nov. 22, 2003]
All our sites are full of content 100's of pages each, all different content and different catagories. Our thoughts were that -va was in fact using a content algo of some sort since we and other competitors ranked well if site contained pages of content.
Another thought too is that google obviously monitors this forum, why not in future before going live don't they use 1 particular datacentre and use our feedback first, isn't that called brainstorming, and in fact using many more people than you could fit in a room! With anything especially on this scale some factors get missed or not thought of, and with the diversity of site builds and seo techniques some algo factors as can be seen do not work when used so generally.
Also the other piece is the idea of rules. Up till now Google has had defined rules & filters etc that have been pretty much deciphered by the SEO community. If you view this update as not having rules it certainly seems to make more sense.
Google give me back my money keyword backlinks that I paid dearly for or stop encouraging SEO people to get inbound links.
The nice thing about this being a filter is all Google has to do is flip a switch and everything goes back to normal.
One of my 3 index pages has improved its ranking somewhat after getting hammered last weekend by Florida. Up until last week, I was regularly ranked at #3,4, or 5 for my index page's target keyword (for 9+ mos). For most of the last week, my page has been down around 460 or so since Florida started. This morning it is about about 250, and the top 10 looked very different than the past few days.
Another index page for a subdomain is showing similar results. Last week, #3. Most of this past week since Flordia around 250, today it is about 180, and the top 10 are very different.
If google is judging a keyword combo, by the entire site
and not by a single page, that may explain why so many
ecomm sites are dropping like a rock.
Look around then. Our main site with around 200 pages was hammered. A smaller site with around 15 kept its position
You know what replaced our main site? A one page site with the search term in its title and H1 and no content but Adsense ads!
I am not convinced that we are looking at a dictionary (lacks the elegance that Google likes so much), but there is certainly something. Great discussions now that all the whining is gone away.
WBF
I don't think the size of the site has to do with anything. I have about 6000 pages and fell 10 pages from the front.
One thing I have seen that is consistent for many different searches is that the results exactly match the directory results if there are any for that given term. This has been mentioned but not really talked about.
If this is the new way they are blending the algo with directory then think of the ramifications. They only dump the DMOZ results a couple times a year lately. So, now it can take six months or longer to get a site listed if you can get it in DMOZ.
Yahoo does the same thing now. When searching there they replace the title tag with their directory listing if you’re in it. I feel this will play a much bigger role once Ink is incorporated thus giving value back to paying to be in their directory.
If Google has decided to put their future in DMOZ, it is a slippery slope. DMOZ is rife with corrupt editors. When you get spam from a DMOZ editor telling you their position there can help yours in Google, then there is a huge problem.
I can only hope this is some sort of algo test for Google and they don’t go solely on their directory listings. If they do, they will not be as strong a player as they have been in the past.
That’s my one update thread post. Back to lurking.
We removed keyword1 keyword2 from the Title, from the H1, and reduced the page density down significantly, only repeated this phrase once on the index page (<5%). Googlebot visited us and today we are showing fresh results with NO improvement in SERPs.
We have only 3 external links to the site that use keyword1 keyword2 in them.. The index page has a PR4.
And for the first time in a year, from a strictly seo point of view, the thing to do now is not get more and more links links links, the thing to do is to get more and more content on pages.
It's all perspective- but this does not seem the case across the board. Adding content has always been an SEO mantra around here.
I get the gist of what you are stating. I can see 'informational' sites getting a boost in more generic, broad phrases like keyword1 keyword2, but [b]not[/b/] under something like 'industry name company'. <-- That has happened in my little neck of the woods. Luckily the OV CTR has been excellent.
WB
[uweb.superlink.net...]
If you need more info, you can PM me as I don't keep tabs on threads like these.
not all keywords just the major ones
How would Google define 'major'? AdWords?
WB
I think agent10 was right to say that Google should use this forum's voice before making this type of index live. This new index will kill many legimate sites and has made the serps for the most part irrelevent.
If Google isn't finished updating then it shouldn't make the results live. Test, test, and test again should be their method. I believe the current situation will hurt Google. People have little patience and will not tolerate such unless search results.
I'm starting to think giving out more links might be a good idea. Directories seem to be ruling the day for "area widgets" at least. Great time to shop for niche directories!
>Look around then. Our main site with around 200 pages was
>hammered. A smaller site with around 15 kept its position
One of our 6000+ pages sites has dropped 3 pages (to page 5 of SERPs) and our one page site has moved from #11 to #3.
Lots of our 10-15 page sites have not moved at all and some have moved up. Several 25-100 page sites have moved up or down or stayed the same.
Size does not appear to matter by itself.
It does cause confounding effects.
For example:
widget widget <country of the widget> (widget being something every site needs unless you do it yourself)
In this example what seems to be prevalent is lists & pay for inclusion lists of widget companies. (Interestingly the customers of these pfi directories are faring better in updateFlorida)
So once you have the authority sites for your keyword, do sites mentioned (doesn't have to be a link) on the authority site rank better?
Tip, you can preface your authority search with "resources for....."
I'm thinking of registering a new site and might think twice about the name now..
[Hey - if my site goes from #3 to #800 they must be silly results ;-) ]
=======================
If this were to be the case, then how about the overly optimized sites / pages with "less popular" keyword phrases? With less popular and non-generic ones, it does not appear that the sites or pages are dumped by Google. IMHO, Google does not impose any filter based upon "on-page" factor. If that keywords are popular and competitive, then filter is applied based upon "on-site" factor.
You aren't looking very hard then. If the above was true, all brand.com sites would get penalized for "brand".
>Look around then. Our main site with around 200 pages was
>hammered. A smaller site with around 15 kept its position
>> One of our 6000+ pages sites has dropped 3 pages (to page 5 of SERPs) and our one page site has moved from #11 to #3.
================
I second that...
>>> Lots of our 10-15 page sites have not moved at all and some have moved up. Several 25-100 page sites have moved up or down or stayed the same.
Size does not appear to matter by itself.
================
Are your targeted keywords generic and super competitive?
The general trend right now does not seem to support that.
The hard work & time I/we all put into our sites to make them relevant and good resources for people (yes so we can profit but that's fair) and remember Google also relies on our sites to attract traffic to their search engine-- and with the snap of a finger they can take it all away from us overnight.
What is the point? Were their search results not relevant?
Now we are all going to have to take the time (invest more money) to rework our sites so that Google likes them again.
And in the end it will be the same relevant sites at the top that have always been there.
I don't undertand.
Should I support a company that does this?
I don't want to have to re-write every word on my sites. I can't afford to. Then again, I can't afford not to.
Yes my sites are relevant for certain searches, but now all traffic is gone. Why? Just because Google suddenly thinks sites that don't use these keywords are somehow more relevant?
I don't understand.
As far as my own personal search engine queries go, I hate to do it but I have already begune searching at alltheweb.com where I know I can get results similar to what I was getting at Google a week ago.
I know some of you are going to say I should not complain because the traffic I was getting from Google was free.
But ask yourself was the traffic really free?
I mean, consider the time and effort that goes into making the sites all nice and pretty for search engines.
Everyone with the work they do on their sites helps to make Google what it is. Without our sites, Google would have no results to show.
But when one player like Google controls so much (with their results also at Yahoo) I think it is questionable for them to be able to just flick a switch that turns out the lights like this across the board.
Many of us commercial sites have employees we pay based upon the traffic level we are expecting. But now the traffic is gone, and we still have top pay these people.
Am I supposed to freak out and start putting in time to frantically rearrange my site? Or am I supposed to hope that this is just temporary?
It just seems at a time like this that a company like Google has way too much power. If they are going to dramatically change results they perhaps should be regulated in some fashion. In other words, maybe they should only be allowed to change the results for a percentage of the search queries being conducted gradually over time, so as to not cripple various businesses instantly overnight by taking away traffic. You know, phase in a new algo over time. Because these search result changes are so drastic and so is the effect of it all.
Sorry for the venting. Good luck everybody.
don't panic, don't change anything - the results look silly - the sensible ones will be back.
Why would they throw away everything they know about determining the topic of a page? If they're going after "over-optimized" pages, that would be swell, but the way to attack that problem is to put a bell curve on keyword density or something. I believe they already stop counting "hits" after a certain number anyway.
I've seen some of our content drop in SERPs, some of it rise, and a big drop in referrals for a lot of search terms we didn't really deserve to show up for. Overall, Google referrals are down about 15%, but we still show up where we ought to.
It's a bit early to tell, but it seems we're getting broader overall coverage, with more search terms appearing in our logs. Is anyone else seeing "more search terms?"
What in the world is going on over there? At this moment, Google is useless to me (and to the searcher, IMHO). Almost every search I do returns DMOZ, Yahoo directories, Google directories, Epinions and KMart. Not exactly focused results if you ask me.
I just wonder about my funky AdWords campaign results and the relationship to the Florida fiasco.
Hmmmm.... November 22 is an appropriate day for conspiracy theories.
Googlebot visited us and today we are showing fresh results with NO improvement in SERPs. We have only 3 external links to the site that use keyword1 keyword2 in them. The index page has a PR4.
Too bad it didn't help. But I congratulate you on this careful, rational test and your concise report of your results.
This would suggest that Google is keeping, along with a compressed version of the entire page (which is used to extract the snippet), a list of keywords from external links to that page.
I had my doubts that this was happening with external links, because it would mean that much more planning went into this filter than otherwise. The external link anchor text is most likely not collected on the fly. That would be too expensive in terms of CPU overhead. It's the sort of thing that would be collected once per crawl -- just like the old PageRank had to be computed once per monthly crawl.
But it makes sense that Google would someday use external links as part of the filter scan. I'd try zapping the one mention of your phrase in the index text, so as to completely sever the connection of that page from the external anchor text.
The reason I say that is because my favorite example of external anchor text making all the difference was that a search for "discount brokers" (don't use the quotes) used to produce an empty directory in the number one or two spot. I checked it three days ago and this was still happening. It's been empty for a year, and Google has been doing this since I first noticed it in April. But I checked just now and it is gone! For the very first time since April!
Maybe Google now requires that the external anchor text appear somewhere on the page before they give it credit. Now that would be, in my book, a real improvement.
Keep in mind that Google can react very quickly with this new filter. There might be some AI (neural networks, self-organizing maps, etc) elements involved, but even so, it seems to me that Google would do most of the "training" of the algo off-line and begin with a fairly stable dictionary. I can't believe that major dictionary shifts would be tolerated online -- it would be too unstable and ruin Google's reputation.
One more thing. There are a lot of comments about keyword1 and keyword2, etc. It would be very helpful if more posters tried to make a determination, while the "-wqwqzw" test is still available to us, whether either or both of the keywords appear to be "dictionary sensitive," either alone or when paired, and when paired, whether they are sensitive to which one is first.
Occam's Razor suggests that all else being equal, you should go with the simplest, least elaborate theory that explains the phenomenon... it seems to me that if non-competitive and informational searches have been very stable while competitive commercial ones have been volatile, the simplest explanation is that one of the previous major optimization factors has been reduced in importance, causing a big shake-up among the most highly optimized searchterms. If it was a penalty, then either all the searches would have been volatile across the board (not true) or Google would have had to spend a tremendous amount of time on this putative dictionary, which wouldn't even account for why some commercial and even spammy results are doing fine for competitive terms. My money's on excessive repetition of keywords simply having been de-weighted somehow, possibly combined with some merging of singular and plural words and the like. That would account for all the industry-specific chaos without elaborate conspiracy theories or incorrect predictions.
I'm sorry this is causing such financial distress for so many people, because it's rather fascinating from a purely observational perspective. )-:
==========================
This Florida update is quite harsh this time for online business as well as of the timing - Many people, both who got bump and got dump, are caught "unprepared"!
For those who got bump from nowhere to top spots get too much traffic, orders, inquiries than they have anticipated and prepared thus it is pretty difficult to handle the incoming volumes - products running out of stocks and employees working overload - thus resulting in lower quality of services to customers who are the endusers.
On the contrary, those who got dump I can imagine the frustation - products in warehouse are left unsold, employees no work to do and so on...
[although I have worries about the Google Razor ;-) ]
Hmmmm.... November 22 is an appropriate day for conspiracy theories.
Heck, Every day is an appropriate day for conspiracy theories on these boards. Even the date of the first Apollo landing will do!
Occam's Razor suggests that all else being equal, you should
GG, 2nd request for a comment (please).
we see NO penalty for keyword1 or keyword2 or keword2 keyword1
However, when we search keyword1 keyword2 keyword3,
a interior page ranks higher than our index page, our index page is indented to our #3 ranking interior page.
Anyone been hit badly on sites that haven't developed new links lately, i.e. messed up worse than pre-September levels?
I'm at 2 & 3 there I'd settle for that , 22-200 elsewhere.
the search for keyword8 <country name> on www-va brings up more than 3,5 million results, me at #2 with a sub-sub-sub page (PR 0). keyword8 is a very very generic term resulting in more than 17 million pages on Google. It's really our 8th keyword as we use it only for descriptive purposes.
The search for keyword1 keyword8 <country name> on the same datacenter has only a total of 33,600 pages. And we cannot be found at all.
Keyword1 is the name of the industry I'm working in. A search for keyword1 shows my site on -va on #80 the index page being a PR6
Both keywords apear in the title of the page, both twice in the body. There's no description for this page.
Does this sound like anyone will like the results? Anyone any questions?
>>> The search for keyword1 keyword8 <country name> on the same datacenter has only a total of 33,600 pages. And we cannot be found at all.
I repeated that with the "-wqwqzw" and my sub sub sub page shows up #1 plus a second much more relevant page from my site. Hope this helps
So today I downloaded that abc tool for checking pagerank across the data centers.
Having inputted the url of my site into the program, the results report "Dea" in every slot per data center, instead of a pagerank value.
I have inputted some of my other sites and it works fine.
Does anyone know the reasons for this? PLZ
As for this google update, being totally truthful some of the results are not quite a bad as some are making out. My sites are far down the serps unfortunately, but once I figure out why this is then I can go forward.
My favourite set of results so far are at [labs.google.com...]
lol - I don't think that set of result counts anything towards the latest update, but the results are completely different :)
That top 10 site for "jewelry" is one of the most un-optimized sites I've ever seen.
I Hope G's Investment Bankers are not reading this stuff. If so, they must be heading for the hills!
-c
Again, NO. My 2 amateur sites happen to have all the above, and I have NOT dropped for the relevant keywords.
Google WHAT ARE YOU DOING?!?!?!?!
This is total madness!
I believe it! My jewelry site vanished after being #2 or #3 and always in the top ten position on a certain two word search for 2yrs. Noticed the other top tens also have dropped drastically or disappeared. Now just junk coming up in the results.
Noticing an increase of other search engines in the referral stats. At least I'm still #2 or #3 in those.
Trying to find good search results now has been a waste of time. I've switched to other search engines myself now. Hope they sort the mess soon.
(a real estate site selling ranches) is top ten for jewelry. Can anyone beat that?
:)
Thats a classic!
Can't really comment on that much.
Check that backlinks and you'll see how it got there. I have a suspicion that the page used to have relevant content for "jewelry".
1) The domain does NOT end in .com
2) The page is in a language other than English
off topic , gotta wonder about Brett's bandwidth bill this month?
Aji
