| 1:54 pm on Dec 18, 2003 (gmt 0)|
Noticed this earlier - allinanchor searches are showing a completely different set of results to yesterday. Currently pretty much in line with a straight keyword search.
If I understand correctly and you're looking at the back links, presumably those links would have shown up there before, but whether they previously carried much weight is another matter.
Either way, assuming this is not a simple glitch or a red herring and it stays like that, this may well be worth looking at closer with regards the Florida update.
| 2:06 pm on Dec 18, 2003 (gmt 0)|
From the example given, maybe it's to do with breaking the keyword/phrase pattern.
| 2:06 pm on Dec 18, 2003 (gmt 0)|
>>including KWs from file names in their allinchor calculations
How's that? I'm not quite getting it. KWs from filenames where?
|2. Just the Url www.*****.com/webmaster-resources-widget.html |
You mean just the URL on the page, without it being a clickable links - just the text on a page, is that what you're seeing being counted as a backlink? If so, I saw it for the first time a couple of months ago. It was the URL written just on the page, no link.
Is it that you're seeing, or something else?
[edited by: Marcia at 2:25 pm (utc) on Dec. 18, 2003]
| 2:22 pm on Dec 18, 2003 (gmt 0)|
>KWs from filenames where?
I guess from the back links, but I'm not 100% sure.
Irrespective of that though check allinanchor searches - yesterday, we were number one for several searches with allinanchor and today nowhere - just like Florida.
| 2:51 pm on Dec 18, 2003 (gmt 0)|
Marcia they were clickable links, but the anchor text was just the company name.
| 3:20 pm on Dec 18, 2003 (gmt 0)|
Sorry glengara - I misunderstood your original post - presumably then, you're talking about outgoing links. Previously, I always thought allinanchor searches related to incoming links only.
| 3:31 pm on Dec 18, 2003 (gmt 0)|
Sorry, to clarify, those were the links returned as pointing to the page that was no.2 on an allinanchor search for "widget resources".
As mentioned only one link had the KWs in link text, the other 3 had the KWs only in the Url itself.
| 4:37 pm on Dec 18, 2003 (gmt 0)|
This would give some more hints that Google is not only looking at the anchor text when deciding what a link is about. Wouldn't it?
| 7:26 pm on Dec 18, 2003 (gmt 0)|
Might even out the playing field a bit, all those recips with optimised link text would have a bit more competition.
Still need to be savvy enough to use KWs in file names though.
| 9:45 pm on Dec 18, 2003 (gmt 0)|
Is everyone assuming that G *likes* the presence of those kw's in the string?
| 11:57 pm on Dec 18, 2003 (gmt 0)|
Google has always liked keywords as file names, directories, etc.
| 12:30 am on Dec 19, 2003 (gmt 0)|
>>Is everyone assuming that G *likes* the presence of those kw's in the string?
| 12:44 am on Dec 19, 2003 (gmt 0)|
allinanchor:keyword returns as #1 a site with exactly two backlinks showing in Google (19 in ATW). Of the 2 in Google, only one has the keyword in the anchor - as part of an 8 word phrase "Company name blue widgets in city city, state"
The anchor text in the other link is..... "website". No BS. It says, "Visit their website."
| 12:49 am on Dec 19, 2003 (gmt 0)|
>allinanchor:keyword returns as #1 a site with exactly two backlinks
And where is that site(page) showing in regular serps?
| 12:54 am on Dec 19, 2003 (gmt 0)|
It is #2
It has been top ten for over a year. It's a page that says "Company Name has been sold (company and widgets) and no longer exists."
It is one of the more relevant pages these days :o
(Not really, a bit of sarcasm there)
| 2:32 am on Dec 19, 2003 (gmt 0)|
> Google has always liked keywords as file names, directories, etc.
Check out GG's message #58, early in Florida, where he pulled out this quote from another WW member, and referenced G's intent to improve the quality of SERP's:
"If this is a new set of results, as opposed to a rollback, then mirror-merchants and keyword-in-link are deader 'n disco..." [WebmasterWorld.com]
| 2:40 am on Dec 19, 2003 (gmt 0)|
Google Guy has said many times that linking with appropriate keywords is just fine.
Mirror sites are getting hammered, thankfully.
Also, "Google has always liked keywords as file names, directories..." isn't talking about anchor text. It's talking about words that would be in a URL, like site.com/keyword.html or site.com/keyword/ or (much less interestingly) site.com/keyword.jpg All these things are still fundamentally sensible webmastering and valued by Google.
| 3:02 am on Dec 19, 2003 (gmt 0)|
|allinanchor searches are showing a completely different set of results to yesterday. Currently pretty much in line with a straight keyword search. |
I don't believe the allinachor results being presented. I know my sector pretty well, and the sites at the top of the allinanchor listings don't have anything like the most in anchor. I wonder what's going on?
edit: allintext results are suspect too.
[edited by: superscript at 3:20 am (utc) on Dec. 19, 2003]
| 3:06 am on Dec 19, 2003 (gmt 0)|
I wasn't talking about anchor text either. I was referring to "keyword-in-link.html" as a potential issue, and following on some of the comments above, like:
...it appeared to me that GG was commenting on this as a potential negative, and with the possibility raised of an OOP situation, this certainly would have to be considered a potential red flag of the sort that G may target, currently and/or in future.
We all know what those "localcity hotel" SERP's looked like pre-Florida. ;-)
Also, I'm not saying it *is* a problem...only that it might be one factor playing a role in a larger picture. When GG makes a comment as overt as the one referenced above, you gotta pay a little bit of attention.
| 3:11 am on Dec 19, 2003 (gmt 0)|
A url is not a link. His comment doesn't even address page-name.html. More to the point, if you have a site or page or section about apples, site.com/apples/ and site.com/apples.html is still straightforward good SEO which has no negative implications and is and should be valued (slightly) by Google.
| 3:41 am on Dec 19, 2003 (gmt 0)|
I agree! A page name is most likley a factor in ranking. It is also very logical and useful to name a page inline with what the page is about. It's VERY doubtfull that it would do any harm, at worst Google probably ignores it.
| 3:50 am on Dec 19, 2003 (gmt 0)|
|More to the point, if you have a site or page or section about apples, site.com/apples/ and site.com/apples.html is still straightforward good SEO which has no negative implications and is and should be valued (slightly) by Google. |
Looking at the SERPs, its pretty clear that this practice is indeed valued, perhaps too much recently. In some cases SERPs are dominated by buy-b*st-widget.com, buy-b*st-gadget.com, buy-b*st-wodget.com - all owned by the same company and using a highly formulaic approach to site and page naming.
| 5:27 am on Dec 19, 2003 (gmt 0)|
It does not look to me like one can trust allinanchor, allintext or allinurl at this point.
Not sure what it all means. Kind of like plugging the -qwerty results. The allin*: results are being brought in line more with the regular SERPS.
The only constant is change.
| 8:50 am on Dec 19, 2003 (gmt 0)|
Indeed, naming files to keywords separated by dashes is valued by G, yet you'd have to be carefull: overdoing it, e.g. using 4-or-more-word filenames might actually decrease G's appreciation.
I also noticed the results constantly changed, as do Scroogle's. Yesterday I pressed the search button 4x in a row, and kept getting different results. A normal G search however, is pretty consistant.
| 2:25 pm on Dec 19, 2003 (gmt 0)|
|A url is not a link. His comment doesn't even address page-name.html. |
Learn something new every day. ;-)
steveb, given that the title of this thread is *Allinanchor and File Names* ... what I was trying to do, obviously not very clearly, was to suggest that kw loaded anchor text in conjunction with keyword loaded file names, is a now a potential concern.
I don't believe that domain.com/keyword.html would ever represent an issue, but green-widgets.com/buy-best-green widgets-here.html, might be.
| 4:36 pm on Dec 19, 2003 (gmt 0)|
If using and repeating KW in folder names is considered spam by Google you'd never know it by looking at my competitors.
They have mycityaccommodation repeated 4 times (in 4 folders) on their way to the final html which is the same as another 80 htmls in the 1st 100 results.
They simply spread them out over different domains and change the background color...
I'd be more than happy to sticky anyone the search page, seeing is believing.
| 7:56 pm on Dec 19, 2003 (gmt 0)|
Regarding caveman's last post, do you or anyone believe that a url, which has keywords in the url, for example, www.kw1-kw2-k3.com, that Google would not like this?! I consider myself a "whitehat". Our domain name is like the above example. It also happens to be the name of our company, what are expertise is, and what we sell.
We've been number #1 for few years, and then post Florida, it essentially disappeared for a search for the combination kw1 kw2 kw3. We are still #1 for a search for kw2 kw3. Does that make any sense?
| 9:04 pm on Dec 19, 2003 (gmt 0)|
rshandy - *my opinion only*:
I do not believe that "keyword-in-url.html" by itself would ever trigger a penalty or filter.
What do I believe, but cannot prove is:
-- G is attacking spam more directly now than ever.
-- There is a filter being applied at the query level and it seems to relate to over optimisation of sites. I am certain that it is *not* just $$ keywords as some have suggested (too much evidence to the contrary). I believe that it's tied to over optimisation, which almost certainly would also cause more carnage in highly competitive, commercial areas, hence the speculation about commercial filters.
-- Too many keywords in the URL string (i.e., file names) *may* be one element that G's algo considers when trying to identify over optimisation.
-- There are numerous other optimisation techniques - that some may consider basic webmastering while others may not - that would necessarily be looked at if the filter exists. Taken together, these elements might generate some type of OOP score.
-- That score could conceivably be triggered at different levels, depending upon the keywords in question (wild speculation).
In any event, I fear that what G introduced with Florida may get worse before it gets better, if it ever gets better, but I'm *truely* hoping I'm wrong.
| 10:13 pm on Dec 19, 2003 (gmt 0)|
"In any event, I fear that what G introduced with Florida may get worse before it gets better, if it ever gets better, but I'm *truely* hoping I'm wrong."
Yeah, I hope you're wrong too...but this does seem like just the beginning. I'm also hoping it can eventually be rectified and that their main objective was to reduce non-germain "spam" sites dominating the top search positions, while still indexing sites truly dedicated to the search subject matter.
| This 35 message thread spans 2 pages: 35 (  2 ) > > |