Forum Moderators: open
But my site still is listed, and i still get hits for all my other search terms, the much less important ones..
Have i been penalised for just that search term? I just dont understand it at all.
Many thanks,
Soapystar mentioned it was a naughty topic.;)
Did fresh tags coincide with your pages demise?
If you're not familiar with fresh tags they are at the bottom of the description on Google. For instance
dmoz.org/ - 17k - Jun 26, 2003 - Cached - Similar pages
where the date is the fresh tag.
<edited for not enough link tos.
What you describe I have seen many times (but not something recent).
Powdork, was that a faux pas, or just a play on words? ;)
Chrisandsarah,
For what it's worth, I've been having the same problems with one of my sites. It relatively new, so I haven't been stressing too much about it. But like you, I get high listings for the search terms I've optimised for, only to have the page vanish (but still indexed) for those terms. I am, however, getting some really obscure hits for terms I'd never even think to target. I don't think it's any kind of a penalty though. It's just Google going thru adolescence.
I've just decided to take the blue pill and accept it as a way of life for the time being. It's a good time to develop other sites in the meantime.
2odd...
This is true. A "selective penalty" has existed for some time. GG may have stepped in up a notch with the recent changes.
All I can say is if you are "manufacturing"/trading for all of your links, you may want to analyze the way in which you are doing this.
Precisely what are the circumstances this happens under? I would presume when all, or almost all, links have the same anchor text. Problem with that idea would be what if I start marketing my newly invented sprockowidget and sprockowidget.com? With the product being called "sprockowidget", it is likely everyone would link to me with that anchor text. The tricky part with such a filter is not dinging innocent sites.
The main site I try and optimize may be suffering from something similar to the keyword penalty described in this thread and I would like to know if it is, and what causes it etc etc
Arn't these valid subjects for discussion here?
Index pages that disappear for a particular keyword search often (frequently, always?) seem to be pretty innocent pages, with no hidden text, spam, keyword stuffing, etc.
At the same time, pages with hidden text, spam, etc. etc. are not disappearing for those keyword searches.
I've worked with databases in the past, and if I did a search for an employee's name, and the employee's records didn't show up, I wouldn't assume that the poor slob had received an employment penalty, or wasn't dressed properly, or had egg on his tie. I would assume that something in the linking structure was screwed up between the index and the database. If I had a distributed database, with data centers all over, I'd be even more suspicious that the indexing, or retrieval, or storage was in error.
Brand me a sceptic!
P.S. GoogleGuy seems to be keeping his head down!
I'm not going to worry about it though, i guess it was a temporary glitch or the everflux?
I think i may give the anchor text in my incoming links a bit of variety though from now on. Just incase
but should sleep better tonight as it's back in second spot!
as Napolean said, (the webmaster one), once you are in the twighlight zone you never come out.
That is definitely part of it.
It is highly unlikely that if a site is about widgets, everyone would "naturally" link to the site using widgets only as anchor text or any one phrase ONLY, without being PROMPTED by the site owner.
Google's algo is designed around the assumption that pages link to other pages as a "vote" and do so naturally and they are trying to make sure that votes are legitimate.
If a site has all their links "buy widgets" and nothing else, the vote has most likely been fixed or tampered with. Most likely, if the links came "naturally", they would look more like:
1.widgets
2. buy widgets
3.widgestonly.com
4. click here
5. widget store, etc...
<<The tricky part with such a filter is not dinging innocent sites. >>
As with all filters, many innocent sites are caught in the net.
A normal set of backlinks would include
Click here
kw1 kw2
mysite
my site
mysite.com
www.mysite.com
[mysite.com...]
[mysite.com...]
Visit mysite.com
And then if your title exactly matches the link text and it's in your h1, things start looking very suspicious.
Has anyone lost their main keyword using a title "mysite.com - keyword"? Is that enough of a difference so that the site does not get dinged?
- anchor text was part of the problem
- trading links wasn't part of the problem
- on-page factors were the major problem
- the problem was solved by changing the the affected page (and nothing else)
Of course, other people could have made other experiences.
One more question: Has anyone lost a main keyword which is part of the domain name? (e.g. www.lostkeyword-keyword.com)
If _many_ innocent sites get caught, then this would be a very lousy filter. Totally avoiding any collateral damage is realistically impossible. However, this being common would not be a good thing.
allinanchor:click here should produce a different set of results (particularly a different #1 listing) than just searching for click here. This is not the case. And the results are uniform for this search across all 9 datacenters.
I believe this is more of a problem with the switch from collecting data from 2 bots down to one, rather than a deliberate change. If it's a change, it's not a very good one.
Okay, so if this theory is correct:
allinanchor:click here should produce a different set of results (particularly a different #1 listing)
No. The anchor text is just a part of the problem.
If you haven't any experience and you draw conclusions, it's just speculation.
(Please post facts, neither theories nor speculation.)
I have examined a dozen of pages/sites which were affected by this problem and all facts indicate that this is a keyword penalty.
By the way those guestbooks spammers are a good example to disprove the theory of a keyword penalty. All of those webmaster sign guestbooks usually with always the same link with the same anchor text. According to our theory they all should receive the keyword penalty but the don't.
Sticky me if you want some examples of guestbook spammers who fulfill those criteria.
It seems to me that in many of the cases where people have reported dropping from the SERPs on their main keywords/phrases, they'd also made recent changes to backlinks or internal linking structure.
I think those changes just aren't always reflected consistently across all the datacenters because they're relatively recent and almost not part of the "even more permanent index" yet. From what I've seen, only pre-Dominic links are 100% solidly accounted for and factored in.
Hmm...interesting argument. If this semi-penalty is in place, then these guestbook spammers would be expected to be hit. These spammers typically have no links worth mentioning beyond guestbooks.
For everyones info, my anchor text is varied but for the page in question, 70% of the incoming links anchor text is the same. You wont find it on any guesbooks though, just partially relevant sites.
thanks for posting some facts.
By the way those guestbooks spammers are a good example to disprove the theory of a keyword penalty.
This doesn't disprove my theory because anchor text is just one part. (The 'same anchor text' was a different theory.)
It seems to me that in many of the cases where people have reported dropping from the SERPs on their main keywords/phrases, they'd also made recent changes to backlinks or internal linking structure.
Of course, there are a number of people who mentioning a drop which has nothing to do with any kind of penalty. However, this penalty exists. (I already mentioned the typical behaviour I'm referring to.)
And what are the things trigger this "penalty" doc_z? Knowing that would mean knowing what to make sure not to do with your site.
As already said, I examined several different sites (just two of them are my own).
As far as I know, the penalty is triggered by anchor text plus on-page factors:
Numerous incoming links which include the keyword as anchor text plus over-optimization for that keyword.
I presume you mean external incoming links? All the internal ones being the same would seem normal. How many of these incoming links are there? If there are only 20 external links to this page, 70% of them being the same wouldn't seem at all odd to me. Particularly if the case were the company was called "Universal Widgets", and these 70% were other sites linking using the company name. However, if you have 400 external links, 70% identical might seem suspicious.
