| 2:26 pm on Nov 25, 2003 (gmt 0)|
I don't like cross posting the same information, but i think my post #367 in this thread might in fact be more relevant here:
Google SEO longterm?, page 25 [webmasterworld.com]
| 4:18 pm on Nov 25, 2003 (gmt 0)|
I lost about 12 index pages on 12 different sites.
I am not making an changes to get them to come back yet because I do not know if this thing is over and this is it.
has anyone said that they are no longer seeing movement and that this is it?
| 4:52 pm on Nov 25, 2003 (gmt 0)|
Interesting, the -mt-tb.cgi addition seems to work for me, whereas -anything else does not. With -mt-tb.cgi added, for my keywords, I basically see the results I saw before this update.
| 5:58 pm on Nov 25, 2003 (gmt 0)|
|the -mt-tb.cgi addition seems to work for me, whereas -anything |
This is because "mt-tb.cgi" are not likely to appear on a page whereas "anything" is rather likely to appear and cause the page to be excluded.
For this to work, the word after the "-" operator has to be exotic.
| 6:34 pm on Nov 25, 2003 (gmt 0)|
If your allinanchor:keyword phase search return your site in top position/position before the update. Then I can conclude that it is the special algo (a filter) that kicks in and stop your site from appearing with the normal search.
I just want to say a story (an opinion). It is not a fact.
You try to rank high with the keyword phase 'free adult pictures' so you exchange links with 100+ different site with ALL anchor text as 'free adult pictures'. You then put the words 'free adult pictures' on your title and you also put it in your H1. Then you thought keyword density would help so you make 'free adult pictures' occured many times in your page, some in bold, some in italic, some underlined, some in ALT, and some even as the blue little link text ...
So google count the possibility of your trying to manipulate google so you can rank high with the search 'free adult pictures'. And the calculation yield 99.9% from all the factors they can take it. So they say they can safely eliminate you from the listing - even if you rank #1 for allinanchor:free adult pictures.
The only reason I can suggest that is not related to any special filter is that you get all your backlinks mainly from the same sites, probably bought links. i.e 1000 links bought entirely from a site. And you have 3000 links bought from 3 different site. Of course you can rank #1 with allinanchor search but that doesn't give you much weight on the normal SERPS.
If your allinanchor:keyword phase search return your site far behind or missing too. Then it is normal. Just get more back links with the suitable anchor text.
Missing index page is NOT NEW. It happens before in June/July. Many people thought google was broken at that time while I always think it was done intentionly. But when the algo rolled up that time, it was more like a beta version where it had created too many problems. So they took it away some day in July. But that didn't means they aren't working on it anymore. So today, it is back in an improved form. Of course, it will be continously improved.
What's Da Reason For It? :) To stop people trying to manipulate the SERPs for top listing too AGGRESIVELY.
You can always think Google is broken and wait for them to fix it so your listing is back. This can save you a lot of work becuase you put the resposibility to Google. :)
However, my advice is, Google isn't broken even if it is broken. Because Google set the rules. If you want to play the game, you play according to their rules. :)
I am not looking for an argument with this post, so you don't need to tell me 'Coca Cala' and 'Pepsi' are still ranking #1 in Google. :) Google have more smart brains then me, they do have more great ways to know if it is legit or manipulated. :)
| 6:38 pm on Nov 25, 2003 (gmt 0)|
Can anyone give me a quick general opinion of the following.
My site dissapperaed last night from the SERPS after 6 years, last 2 years first page results all fine..
Now when I do a search for the domain with www or without etc it does not exist.. I get the following...
Sorry, no information is available for the URL myurl
Find web pages that contain the term "myurl"
If I do a search for a non existant domain i.e www.sdfjashdgkasdasgdasd.com I see the following
Sorry, no information is available for the URL www.sdfjashdgkasdasgdasd.com
If the URL is valid, try visiting that web page by clicking on the following link: www.sdfjashdgkasdasgdasd.com
Find web pages that contain the term "sdfjashdgkasdasgdasd.com"
So there is a difference I can see here
When I use the toolbar I still have a PR of 5 and when I use the toolbar to check for backlinks I still have over 150 backlinks...
Any comments appreciated....hoping its just the latest Florida glitch
| 6:53 pm on Nov 25, 2003 (gmt 0)|
I don't think I said SEO penalty, I said penalty for exact matches. I do think exact matches are determined by factors which are associated with what was formerly good optomization. Having keywords in heading, title, body, inbound anchor, with good relevant content, and easy navigation. Thusly the mods have been generous. Please indulge.
tahoe weddings -florida
lake tahoe weddings
lake tahoe weddings -florida -massacre
Without going in too the quality of either set, you should reach the conclusion that the serps when including -** represents a set of results where exact matches are allowed, and even rewarded. This set of results is very close to pre florida (except for no indented results) and what I expected things to look like post florida. Only two of the pages in the first ten listings can be found in the top 100 real results (most likely not in the top 300). Its not an SEO penalty, the sites in the top ten now are seo'd too. If you really want to have some fun look at how many sites in the top 50 are there only because of a link to one of the missing sites.
Oh yeah. And then don't forget to check AV.
| 7:20 pm on Nov 25, 2003 (gmt 0)|
Dirkz - that is not it... when I use -mt-tb.cgi, I see pre-update results for my main keyword term (with my site as #1). When I use -asdfaswerw (or any other random characters), I get very different results, and my site is no where to be found. At least for my site, the -mt-tb.cgi seems to make a difference that random strings do not. Not sure what this means- I have never even heard of Trackback until this thread.
| 7:29 pm on Nov 25, 2003 (gmt 0)|
|If you really want to have some fun look at how many sites in the top 50 are there only because of a link to one of the missing sites. |
that may be the most telling (and disturbing) *feature* of this update yet.
| 7:30 pm on Nov 25, 2003 (gmt 0)|
lorel, you can access different datacentres by using www-ex.google.com. www-in.google.com etc.
Wrt to people finding their pages at #80 or whatever, yes they're actually looking through many serps to find it.
| 7:31 pm on Nov 25, 2003 (gmt 0)|
need3lives. For mine, if it is a two word term -afshfkg will work. If its a three word term you have to add another -jhjhre (or any other jibberish).
| 7:37 pm on Nov 25, 2003 (gmt 0)|
Two MIA index pages. Bloomin' nightmare this is! Of course, using the -scroogle trick they are back there, waving at me. Little sods.
| 7:46 pm on Nov 25, 2003 (gmt 0)|
I have a relatively new site, only about 3 - 4weeks old. It has already been thoroughly spidered by Googlebot and been included into the index.
My main pages and product pages are all turning up and performing really well, but my front page has gone AWOL. Nowhere to be seen. This site has a stellar domain, super clean html and uses nothing but white hat SEO. (Good titles and tags etc)
A few months ago I released another site with a great domain name, same white hat techniques. This site's front page shows up and does really well - number 2 for keyword keyword.
It's frustrating because I would be doing so well if it wasn't for this problem. Sticky me if you want to compare notes.
| 8:08 pm on Nov 25, 2003 (gmt 0)|
I had the same thing happen to two of my sites. It has been like this since June. I tried reinclusion requests but I get no response. I think your sites like mine have been dropped. Usually, they get picked up again but it can take many months.
| 8:26 pm on Nov 25, 2003 (gmt 0)|
allanp73 thanks for the info....did yours still show PR and backlinks
Can you think of any reasons why they get dropped.....I'm just gobsmacked.....fair enough if there is a reason I'd accept it but I cant think of one...
The funny thing is my sales are higher today than they have been for the last 6 days.....very strange :)
| 8:42 pm on Nov 25, 2003 (gmt 0)|
AthalonInside, I don't think it's 99.9%, I think what they are doing is a lot cruder than you suggest.
I just tried an experiment to test an idea out, and I hit paydirt!
I was able to change one single parameter, and selectively activate/disable this penalty.
It seems that -
Some of search words in title = OK
Too many of search words in title = BAD
Now, this is likely only part of the picture, and may not apply in all cases.
But hey, it's a start.
More importantly, it lets us observe one of the things that "triggers" this!
Here was the test -
Test #1 -
4 word search - key1 key2 key3 key4
all 4 words present in body
key1 key2 present at start of title
Result = Top-10, similar to pre-Florida results
identical to #1, but added two (-dasjd -hdasu) to
search to see difference.
Result = IDENTICAL ranking to #1
Conclusion = filter not active for this search/page
Test #3 -
4 word search - key1 key2 key5 key4
(changed 3rd word from Test #1)
identical conditions to Test #1,
ONLY difference is key5 also in title
Result = Dropped from Top-10 (pre-Florida) to page 5
Test #4 -
identical to #3, but added two (-dasjd -hdasu) to
search to see difference.
Result = Back in Top-10, similar to pre-Florida results
Conclusion = filter activated by Test #3
So: With ONLY difference being one more word in search
matching title contents, made difference in activating
rank drop vs getting pre-Florida results.
Note that the test page was NOT optimized, no H tags,
no content adjustments, no stuffing, not really any
white hat methods (let alone black), all it had was
descriptive META tags, ok PR, and a good title.
To have just keywords present in body and 3 out of 4
in title to be sufficient to trigger a rank drop would
explain why the search quality is so funky right now!
| 8:46 pm on Nov 25, 2003 (gmt 0)|
That is an interesting theory. The situation you described about placing the keyword phrase in too many obvious places like the Title, H1, H2, Bold and link text from only a few sites. And that Google penalizes it.
Some of my sites, in a similar situation as you described (Only I do not use a lot of backward links from one site.), dropped big time since this weekend. The top listings appear again if I use 'allinanchor:keyword phrase'.
So I am very curious what other members think about this theory.
I would like to invite more members to react and give their opinion about this theory?
[edited by: tribal at 9:24 pm (utc) on Nov. 25, 2003]
| 9:10 pm on Nov 25, 2003 (gmt 0)|
I did a quick test to see if it made any difference - heres what I did:
I replaced my keyword rich title with just my company name
I replaced my H1 tag with an H3 tag.
I waited 2 days for freshbot to update - No change whats so ever.
Obviously this wasn't the most controlled experiment, but it tells me its a little more complicated then just having title and H1 tags that are too keyword rich.
| 9:26 pm on Nov 25, 2003 (gmt 0)|
maybe I should add that the keyword phrase in the link text, title and H1 was 100% the same.
The text in H2 and B contained the exact keyword phrase and more words to form a complete sentense.
I'm running soms tests and hopefully I can add some results within a few days!
| 9:36 pm on Nov 25, 2003 (gmt 0)|
KevinC, that's why I tried to do a controlled experiment. In my case, there wasn't even any H1 (or H3) tags on the page.
I am sure that there are several different factors that can set it off. It is possible there are other aspects of your site which are still triggering this.
However, in my case, I can at least confirm that -
1) It is possible to trigger this by an increase in the # of search words matching in the title.
2) That this is search-specific. For a given page, some searches will trigger it, amd some will not. (This is likely the basis for GG suggesting we check our server logs and see what searches we are getting Google referrals on.)
3) That a particular page is not necessarily "branded" as bad, as it depends on which search you do (In this same context, it may be that a page dropped to obscurity, not for all searches, but paradoxically, only for those searches you optimized it for!)
| 9:44 pm on Nov 25, 2003 (gmt 0)|
I'll play along.
tribal> I would like to invite more members to react and give their opinion about this theory?
It's a good fit for what I see. A page of mine that ranked well pre-florida for "kw1 kw2" had a title of "myname kw1 kw2" and was linked from every page on my site as "kw1 kw2". The H1 on the page is "kw3 kw1 kw2" and I optimized for the phrase "kw1 kw2" in general.
Post florida this page comes up:
#117 for the query kw1 kw2 unmodified (previously my best keyphrase as you may guess)
#5 for allinanchor:kw1 kw2
#4 for allinanchor:"kw1 kw2"
#4 for kw1 kw2 -nonsense
This seems to support the tribal hypothesis pretty well.
The page is an index of articles on the topic, so the optimization is natural enough since the articles are organized around the topic.
Anyhow, I'll change my title and see how that affects this page. It still gets page one results for kw1 kw2ing, which has taken over as my new best keyphrase.
I also noted G is wicked slow to return the results of most any query with a modifier.
| 9:56 pm on Nov 25, 2003 (gmt 0)|
KevinC, tribal, linear, rather than changing your title and waiting to see what happens (that takes time), why not just try different searches with what you have now?
It should be possible to pop the filter on and off, depending on how close the search phrase matches what is in your titles, H1s, etc! Try using some words present in your page but which you did not optimize for.
You should be able to tell whether you tripped the filter on not for a given search phrase by comparing the results with/without some -'s added to the end.
| 9:58 pm on Nov 25, 2003 (gmt 0)|
edit: deleted, didn't think it was significant after all.
[edited by: claus at 11:19 pm (utc) on Nov. 25, 2003]
| 10:06 pm on Nov 25, 2003 (gmt 0)|
If my site is tripping the filter or not is not the questions - It undoubtly is.
I had 3 #1 "kw1 kw2" money phrases disapear. 2 of the three keywords were never in my title or H1 to begin with. So the title and H1 can't be tripping the filter for at least 2 of the keyword combinations.
That being said I have moved up for some phrases - I just can't figure out how to get back with the ones I lost.
| 10:07 pm on Nov 25, 2003 (gmt 0)|
|You should be able to tell whether you tripped the filter on not for a given search phrase by comparing the results with/without some -'s added to the end. |
I sort of thought I had already done that, aspdesigner. I have a high degree of confidence that's what's going on, a filter. I'd like to modify my page to avoid it, because users are going to use the obvious keyphrase query.
Googlebot spidered my page in question just after I posted above and before I had a chance to tweak my title. :( But that page is getting spidered daily (adding to my bemusement over being on page 12)
| 10:35 pm on Nov 25, 2003 (gmt 0)|
To make sure I understand what some of you are saying,...you are suggesting that if we currently have the following:
title: Product Name - keyword1, keyword2, keyword3
Desc.: blah blah keyword1, keyword2, keyword3 blah blah blah
H1: keyword1, keyword2, keyword3 with different blah blah blah than the description
image ALT tages: keyword1, keyword2, keyword3 with blah blah blah
Change the title? Or all instances where the keyword phrase is matched in any of the above and let the page content talk for itself.?
If keyword1 and keyword2 are the filtered keywords when paired together, then just avoid them together or remove them from the title, H1, and Description?
I do know that keyword1 and 2 for my particular 3 word phrase is the penalty and when I do K1 K2 K3 -dsf -sdf we are in top 10 again, without, we are post florida not in the top 1000. However, 2 of the 3 keywords do not describe our product,...it takes all 3.
| 10:40 pm on Nov 25, 2003 (gmt 0)|
<my bad, paragraph removed>
However, these results continue to focus on sites that are far and away higher quality than previously. People who look for "products" instead of information (including information on products) might be disappointed, but that is just because they are a bit behind the times.
New results moved over to .de, .it, .fr and .es about fifteen hours ago, and the difference in quality is staggering. For my niche (where I don't need to know the languages to recognize the quality) the quality is so much superior now that it is hard to describe. In those languages, for one search term alone, removed were about two hundred bits of pure duplicate spam, plus several dozen other bits of piffle. In their place, and ranking higher, are solid, clean sites with good or excellent information on the topic. I did notice that at least one site that should be listed was not listed, and when I looked, sure enough it had a duplicate content issue.
The trashy anchor text results are gone. Wonderful. Using one of those verboten dance tools, I'm getting a kick out of comparing the current results to the old data using the -ustsrs exclusions. The current results are far better, period, if only because the weakest results among the current results are simply low relevance, while the weakest results among the old were pure no-content spam -- sites pretending to be about the topic but really just offering affiliate links to parent sites that might not even be on the topic.
[edited by: steveb at 11:18 pm (utc) on Nov. 25, 2003]
| 10:43 pm on Nov 25, 2003 (gmt 0)|
If my site is tripping the filter or not is not the questions - It undoubtly is.
No, KevinC, linear, I don't think you understand what I'm getting at here. Different keywords searches with trip or not trip the filter for your page. Right now. No spidering required.
Some search phrases will trip it. Some will not. You don't need to change anything on your page to do this.
The reason I suggested the -gfdsf thing was as a way to tell which searches trip and do not trip the filter for your page right now.
Once you know what is setting it off and what doesn't, you'll have a better idea how to adjust your page accordingly.
I just tried another test with a random site, and it also seems to support my idea -
* I did two searches using the same highly competitive 2-word search, one with some -'s, one without.
* I selected a site that got "hit" by the filter, i.e. - it was Top-10 in the "-" search, but dropped in the regular (Florida) search. The title also began with the exact search phrase.
* I pulled-up their site, and looked for non-optimized words on their page I could search on.
* I came up with a search with mostly random body words and only 1 word in their title, and for which they they showed-up. (i.e. - a search they were NOT optimized for)
* I tried that search with some -'s, to see if they were still tripping the filter. They were not - the rank was INDENTICAL with or without the -'s!
So I was able to see 1 search where they triggered the filter, and another search where they did not, for THE EXACT SAME PAGE on someone else's site!
DO you see what I am getting at here? The filter is getting switched on or off "on the fly" when the search is made, based on the combination of BOTH the contents of your page AND the specific search that was entered!
| 10:45 pm on Nov 25, 2003 (gmt 0)|
Steveb, what news story?
| 10:52 pm on Nov 25, 2003 (gmt 0)|
This is an interseting theory and straight from Google.
Read Googles Terms of Service
This is an interesting quote.....
"You may not use the Google Services to sell a product or service, or to increase traffic to your Web site for commercial reasons, such as advertising sales. "
I guess that means if you want to sell products on Google your only avenue is Adwords?
| 10:54 pm on Nov 25, 2003 (gmt 0)|
>However, these results continue to focus on sites that are far and away higher quality than previously. People who look for "products" instead of information (including information on products) might be disappointed, but that is just because they are a bit behind the times.
On a bunch of SERPs I tried at random, whenever I entered "buy widgets", replacing widget with a real product, I got SERPs with people selling them. Searchers shouldn't be disappointed if "widgets" pulls up a lot of sites with just information on widgets. That info sites tend to come up higher on broader searches is actually *expected* with the Google algo. Non-commercial webmasters tend to link to related sites; merchants don't like to link to competitors. Thus, the info sites tend to have more PR and anchor text.