Forum Moderators: open

Message Too Old, No Replies

keyword penalty/bug....google behaviour

how google see's the page

         

soapystar

11:26 am on Jun 29, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Following the idea of a seo/keywords penalty heres something i havent seen discussed and i'd like to know if my observations are shared by anyone else. On pages hit with the bug/penalty i notice that google will ignore exact matches for the phrase but look at seperate occurrences of the words. For example:
a search for keyword1 keyword2 keyword3 will bring up a page but highlight seperate places on the page where the words appear individually and ignore the phrase where it sits as a complete phrase. Before dominic the same search would highlight the phrase as a whole on the page when presenting the serps with the searched words highlighted.

ciml

6:07 pm on Jul 2, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Interesting observation, soapystar. I don't see it where I see specific index pages lower than expected, but I tend to think of it less as "the bug/penalty" and more as "a set of filters".

2_much

6:22 pm on Jul 2, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Could it be anchor text over optimization?

soapystar

6:25 pm on Jul 2, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



well theres no doubt for me that there is a connection. Today all my affected sites are back to normal and now the highlighted text on the serps shows the complete phrase. When the filter/penalty/bug is in affect only seperated words are highlighted.

AthlonInside

6:36 pm on Jul 2, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



how about a 1 word search term?

soapystar

8:01 pm on Jul 2, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



goood point. If its one word the affected page wont show at all on the ones i check. Multiple words will show but lower down the serps seemingly because its giving weight to seperated words but now of course there weighted much lower.

Napoleon

8:28 pm on Jul 2, 2003 (gmt 0)



That's a very good observation SoapyStar... very nice, and very useful. It fits the picture very well indeed and is consistent with a filter.

GrinninGordon

12:28 am on Jul 3, 2003 (gmt 0)



Hi soapystar

Maybe you have something. So having keyword1 keyword2 in that order on your page several times is not good for that exact same search - is that what you are saying? What you need is this keyword2 is a byproduct of keyword1 (made by Peter Smith Inc) but keyword2 is manufactured by Joe Blow Inc.?

DavidT

1:05 am on Jul 3, 2003 (gmt 0)

10+ Year Member



And how about if you have a keyword in the name of your company/site? Do I ask people to link to me with "click here" as the link text instead of site name?

GrinninGordon

1:10 am on Jul 3, 2003 (gmt 0)



DavidT

I think these www.keyword.ext domain sites escape the ne for keyword searches from what I have seen. There is only one question for me. Is the problem with sites a penalty or a data / PR and keyword relevance calculation catch up issue?

MHes

1:15 am on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Hi DavidT

If you have a company/site name which is a non competitive keyword then you have no problem, even if a filter is tripped.

This filter IMHO is only triggered by a series of obvious seo tactics.

GrinninGordon

1:19 am on Jul 3, 2003 (gmt 0)



Hi MHes

Can you say what those tactics are? As I have tried to determine if it could be this based on several search term returns, and am still not convinced.

crobb305

1:33 am on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



There is some speculation that there may be a new anchor text over-optimization penalty. I am not sure if this is true, but I certainly hope not. When people link to my site without my permission, they often just use the title of my index page as the anchor text. Over time, as more and more people link to me, Google might view this as "over optimization" since many of my incoming links would have the same anchor text. Further, GoogleGuy has always said that external factors such as these would not negatively affect a site's rankings or PR because competetors could easily sabotoge.

GrinninGordon

1:44 am on Jul 3, 2003 (gmt 0)



crobb305

Submit www.google.com to 3000 guest books in 1 day and use the title / name "Google". Then, see what happens when you search for "Google" on google.com in 2 days ;-)

DavidT

1:46 am on Jul 3, 2003 (gmt 0)

10+ Year Member



Thanks for replying.

The name of the site/company includes the name of the country the company and I am in. It is a very important keyword in serp terms. Though, in the domain name it is unimportantwordkeyword.com.

On some datacenters, depending on whenever I check, and I think depending on whether index page has a fresh tag, the index page is being buried on the main two word phrases that include the name of the country/nationality.

Without the fresh tag it ranks on first page for most of these phrases. Site is almost a year old and heavily seo'd but all on topic.

I understand that this happening to some other people and trying not to let it drive me up the wall.

crobb305

1:48 am on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



GrinninGordon,

HAHA ;) ...I imagine Google has conveniently excluded itself from penalties.

kstprod

1:54 am on Jul 3, 2003 (gmt 0)

10+ Year Member



MHes...
This filter IMHO is only triggered by a series of obvious seo tactics.

If you mean "against the rules" tactics, then there is no way that can be possible. Too many sites are being affected that do NOT employ such tactics, including my own. If this is what is happening, then Google is having serious difficulties distinguishing between the two.

If you mean normal and approved SEO tactics, then Google needs to come forth and tell us the new rules, rather than punish webmasters for doing what they learned here [webmasterworld.com] and other well respected resources.

GrinninGordon...

Is the problem with sites a penalty or a data / PR and keyword relevance calculation catch up issue?

I would like to agree with the latter. I am hoping that this is the case, as I haven't done anything to deserve a "penalty", and I'm sure some others haven't as well.

goldmund

2:10 am on Jul 3, 2003 (gmt 0)

10+ Year Member



Maybe we should create a table (publicly available) with important data describing 20 or 30 sites affected by this problem and then try to make some useful conclusions out of it.

Napoleon, you have some experience with this, don't you?

But of course, if it is true that it's just a glitch and not an algo change, we'd just be wasting our time......

Chris_D

2:45 am on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I just tested this theory on some of the highly competitive 2 phrase search terms. You know what I mean - the ones where there are over several million results, topics encompassing most of the 7 deadly sins... the highly lucrative areas - the spammy side of the street.

My conclusion is that there is a problem with Google and you are wasting your time tring to analyse it. Its not operating to the published specification - there is no point trying to understand why there are some 'anomolies' in some of your searches.

Try a search for 'free keyword' where keyword is anything that you think may be a 'popular' search term....

Now. Look closely at the serps. Look closely at the cache. How many are just plain badly spammy? Most of them. Spam still rules where it pays - even blatant machine generated stuff - javascript redirections - and even hidden text/ hidden links. On less competitive search phrases - you aren't seeing how bad it really is. Go and have a look.

Google isn't operating to the published specification - its broken - and I think its a waste of time to try to analyse it.

Just hope they sort it out soon.

Zapatista

3:18 am on Jul 3, 2003 (gmt 0)



They finally sorted it out for one of my sites.

Let me explain. 2 months ago, at the start of Dominic, one of my top ranking commercial websites (no spam, optimized in a search engine friendly manner, with mixed anchor text) disappeared completely. Poof. Gone.

I haven't seen it again in the top serps for 2 months. It didn't go in and out like some sites, it was gone and I never saw it again until 5 minutes ago.

For the first time in 2 months, it has showed up at:

www-ab, www-in, and www-fi and www3.

This is a big deal because unlike some of my other sites, it never went "in and out" of the top serps. It left and never came back until tonight.

I hope, I just pray, this is the start of their calculating ALL of the algo factors.

This site was optimized for 2 sets of keywords (Affordable Widgets + Widget Discount Center) and the h1 was taken out and replaced with a logo. I also moved the site over to a static IP and rebuilt the website.

I am also seeing other websites that once ranked high then disappeared. They have come back as well.

Napoleon

12:25 pm on Jul 3, 2003 (gmt 0)



>> Napoleon, you have some experience with this, don't you? <<

Yes. Basically that's what I have been trying to do here (not in public anymore though because I felt frowned upon by some mods).

I basically sacrificed my sites, plus a couple people have handed to me in sheer desparation, to experiment with (as posted in another thread I could no longer sit and wait and assume that everything would be OK).

Essentially I have made different changes to each site affected. This includes anchor text where I could get at it or influence it.

It's now sit back and wait time - there's not really much else I can do except wait and see what happens. And hope.

The fluctuation is of course the problem. If this was a straight filter or rule it would reflect across all centers and certainly would be consistant at each individual center. It isn't.

This therefore suggests that: it isn't a filter; or it is a filter that Fresh ignores (or vice versa); or it is a filter that someone can't keep their hands off.

I have my views on this, but really, for certainty, we will have to see what happens in the coming days. Or of course GoogleGuy could comment (ok ok... I'm only saying!).

Iguana

1:11 pm on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



I thought I was pretty immune to this sort of speculation - I've seen very little change in my pages SERPS for 6 months. However I suspect that there may be a excessive anchor text filter.

I put a panel on a thousand of my pages with links to my 'most popular pages' - mainly to see if I could influence the SERPS though the number of internal anchor text links. These pages have dropped considerably (e.g. 160 places) in searches for the anchor text - but still remain no 1 for the full phrase they are targetted for.

Marval

1:37 pm on Jul 3, 2003 (gmt 0)

10+ Year Member



Chris D...you have a very good point since Im one of those sites ...and have been experiencing that flux...happened again last night (just about a weekly occurence now) and watched as 5 datacenters showed original results, and the others showed a new batch...only difference this time is there are no fresh tags showing up as there have been every other time...
Gotta comment on something else people have been stating matter of factly...if spam filters are being applied, they arent set up correctly as none of the items GG mentioned in his "answers" are being affected....Im of the opinion that they have not started any penalties for keywords and I sure wouldnt be changing any pages based on that supposition.

I am still of the opinion that since they have not done a complete deep crawl, they are working with incomplete results, some dating back to February, and piling freshie results on top of that. When this update finishes we will see if there are any penalties to be applied.

doc_z

1:41 pm on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



The fluctuation is of course the problem. If this was a straight filter or rule it would reflect across all centers and certainly would be consistant at each individual center. It isn't.

There is another explanation for the fluctuations. Google have different versions of the same page chached (As far as I know, before the freshdeepbot area, one was the cache from freshbot and one from deepbot). Therefore, if one if these versions if causing a problem and the other doesn't, this could lead to some strange behaviour. If this would be true, it could explain some 'in and out'. Also, if this would be true, it would indicate that these problems are caused by on-page factors (due to the changes between the different versions).

Can someone confirm or exclude this? (e.g. if someone observe large fluctuations, but haven't changed that page for months.)

martinibuster

1:43 pm on Jul 3, 2003 (gmt 0)

WebmasterWorld Administrator 10+ Year Member Top Contributors Of The Month



This filter IMHO is only triggered by a series of obvious seo tactics.

excessive anchor text filter

It's my own belief/observation that it's possible to over-optimize and look unnatural to Google. Some of my client's competitors try so hard, with the most obvious "optimizations"- the only apt comparison I can make is that these sites look like they have Fake Boobs.

There's nothing natural going on. At a certain point the optimization stops being about "clarification" of the message, and it becomes something not natural, perhaps something that Google isn't looking for.

Like, yeah there around 100 Google algo factors, but did you ever stop to think that Google has studied sites that hit too many of the factors and realized there's a threshold at which no natural site would score on the 100 factors, a threshold at which only an unnatural site would score?

With that said, I know of one competitor who has the hallways and doorways and dozens of anchor keyword optimized text links at the bottom of every page (doesn't look like inbound anchors are optimized, though), and it's consistently in the number one spot for certain keywords- whereas my client's natural looking site occupies around 5 positions in the top ten. It's like Google preferred the natural and gave the site extra positions, but only gave one position to the Silicone Site, and less coverage overall with a broader set of keywords.

Just my 2 pennies.

Marval

2:06 pm on Jul 3, 2003 (gmt 0)

10+ Year Member



doc_z I can definitely exclude this..I posted a few weeks ago about one site I have that was seeing the in and out for some results and not others...the only difference I could find as the page had not changed, the cached page was exactly the same page (I do rotate a text ad that lets me track their copy) however the reported size of the page in the SERPs was different by 4k. The size that they were reporting incorrectly was the size of the page during the worm attack deepcrawl back in Feb/Mar timeframe.

martinibuster...I would have to wait until we see the rest of the backlinks get brought in to make a judgement on the anchor text...some sites are still missing hundreds of links from good authority sites (again this problem started back in the bad deepcrawl) so its kind of hard to tell the effect of the anchor text yet. Im making an assumption here that backlinks that have similar anchor text links are not being penalized since this would affect too many of the portals etc.

MHes

4:52 pm on Jul 3, 2003 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Hi kstprod

The main problem is knowing which algo is effecting whom. Sites drop for many reasons, a clean site dropping does not mean this algo is not possible, it merely means there is some other reason for that particular case.

Normal and approved SEO tactics as outlined by google are fine, but unnatural bending of those rules is not. You can use a hammer to put in a nail, or use it to break into a house.... same perfectly legal tool, but you get in trouble with the police algo for one of them :)

Martinbuster clarified my thinking perfectly. If you over do the seo and manipulate your text accordingly..... beware!

martinibuster

4:57 pm on Jul 3, 2003 (gmt 0)

WebmasterWorld Administrator 10+ Year Member Top Contributors Of The Month



but did you ever stop to think that Google has studied sites that hit too many of the factors and realized

That should read, "...that Google has studied probably studied sites..."

Catnip

7:22 pm on Jul 3, 2003 (gmt 0)

10+ Year Member



Ok well I never thought (prayed) I would have to add my comments/story to this topic... but the time has come. Website is a PR6 and has been on page 1 for our main keyword "blue widgets" (travel related) for the past 6-8 months. And it just disappeared two days ago and has never been seen since. The 2 phrase keyword returns 2,380,000. results, so it is a competitive keyword. about 10 days ago we changed some info on our main page. I raised the keyword density from 7.5% to about 11% and that is the only thing I changed. Never touched the title <h1> tags, etc.. just slightly increased the density and the site moved 2 days later to #8 from #10 and stayed there for about 4 days and poof it has been gone completely now for about 3 days. And I can say 100% that we have NEVER used any dirty/sketchy tricks on our website. Maybe I have missed something on these 10,000 threads, but I think this just calls for a few beers and time to see what is really happening. But any insight or speculation is welcome.

Thanks,
Catnip

More Traffic Please

7:25 pm on Jul 3, 2003 (gmt 0)

10+ Year Member



"My conclusion is that there is a problem with Google and you are wasting your time tring to analyse it."

I could not agree more.

For every "natural" unSEO'd site that has moved up in the SERPS, there is a blatent spam site that has also moved up as well as traditionally well SEO'd site that has also moved up. Then there is the ridicolus. I have held the #1 position for a particular phrase for over a year. I am still first page, but I'm beibg beaten by the Navigation frame of a framed site. It has no text, no alt, even the links are images and javascript. Google is committed to returning the most revelent serps they possibly can. There is no reason a well established site should be #1 one day, #435 the next day, #23 the following day etc. There does seem to be a fairly clear pattern of issues that newer sites are having. Other than that pattern, trying draw conclusions on the serps in general is an act of futility. For the time being, Google is broke. I'm not touching my site until the data centers remain calm for at least a couple of weeks and some patterns start to emerge.

This 83 message thread spans 3 pages: 83