What Do Google's Allinanchor, Allintitle & Allinurl Results Tell Us?

[bobmutch]

MODERATOR NOTE: Most of this discussion was split from Repeating Terms In URL Helpful? (http://forums.searchenginewatch.com/showthread.php?t=1505). In addition, Bob later started a separate thread on the topic. All discussions have been combined into this one.
My position is that allinurl:keyword, allintitle:keyword and allinanchor:keyword are commands that return results showing what pages in the index have the highest Ranking weight for keyword in the url, keyword in the title and keyword in the anchor text.

I brought this up in another thread and there was a couple of people that said they had never heard of this position before. I would like to hear from SEO experts on this position where it is correct or not.

If you do a search on web design you will get results back with webdesign.about.com is bolded. But if you do a allinurl:web design you will find that the results all contain web-design or /web/design/ but you will not find in the results webdesign (unless there is another web and design in the URL). This shows that when you search for web design and webdesign is bolded you only get allinurl Ranking weight from web-design or /web/design/ type URLs and not from /webdesign/ .

[Mel]

I think it is a bit of a stretch of the imagination to assume that just because Google chooses to bold a word in the SERP display, that this is somehow an indication that reveals how the Google algo is ranking such terms.

[bobmutch]

Everyone: What the bolded words show is the words that an allinurl:web design search based the results positions on. It is not the Google Ranking algo that is being discussed but the allinurl: command.
The bolded works do how ever indicate the words that the allinurl: command is useing to determine the results position for that command.

[Mel]

Everyone: What the bolded words show is the words that an allinurl:web design search based the results positions on. It is not the Google Ranking algo that is being discussed but the allinurl: command.
The bolded works do how ever indicate the words that the allinurl: command is useing to determine the results position for that command.

Bob can you offer us any evidence that the way Google chooses to display its SERPs is in fact indicative of the words used in the ranking process of the allinurl:search?

For myself I simply do not think that is the case, but would be willing to be convinced otherwise.

[bobmutch]

No I don't have any evidence concerning the weight that keywords in the URL play. It is gererally accepted that keywords in the URL have Ranking weight - how much, no one really knows.

How ever I can offer evidence, and just have, concerning the command allinurl: and what form of keywords it uses in determining the results position from the allinurl:keyword command. It is the latter, not the former that my posts have been discussing.

Just do an allinurl:web design, or any other 2 work phrase that suites your fancy, and you will quickly see that the results are made up of the key words web and design seperated by a dash or forward slash or as a file name or directory by themselves. In most cases you will not see in the results entries that show the two words together (i.e. webdesign) in the URL.

[Mel]

Sorry Bob that should have been allinachor:URl but I was mostly referring to the idea that you can look at the bolded words in the serps and make the assumption that if they are bolded then that means that Google used them in the rankings of the search.

What the allinanchor:web design vs. all in anchor:webdesign IMO shows is that Google does not recognize run together words as seperate words, and that is true in any of its searches, it is not unique to allinanchor: searches.

A search for
webdesign returns 19,100,00 results
web design returns 28,100,000
"web design" returns 15,100, 000 results.

While these are estimated numbers of pages and may not be correct, they do show that each is considered as seperate terms.

[bobmutch]

"I was mostly referring to the idea that you can look at the bolded words in the serps and make the assumption that if they are bolded then that means that Google used them in the rankings of the search..." Well they are only bolded in the allinurl:web design search, and yes I am saying that the bolded words in are used in the ranking of the allinurl:web design search.

This discussion is not about allinanchor: but allinurl: as my 2 former posts show. This discuss has to do with keywords in the URL and nothing to do with key works in the anchor text.

[Mel]

Ok Bob so to get back to the point what is your evidence that the bolding of words in the allinurl: search reveals the way that google is ranking for those keywords?

If you search for a term in google using any of the various forms of search it will of course be ranking on those words, and you can clearly demonstrate that with a search or two.

My question why and what evidence can you show us that bolded words are anything but a display convention?

[Anthony Parsons]

Bob, I would also like to see your evidence. I have tested this theory and found the exact opposite. To begin with, an allinurl query has nothing to do with rankings. Searchers do not use an allinurl when searching. You might, but 99.99% of searchers do not.

If we are talking Google only here, then here are some facts for you. Test yourself if you like. The bolding of terms is nothing more than a representative that the word was found. Google DOES NOT use the description for ranking purposes, though will bold the words found within the description if used instead of a snippet.

As David said earlier, names in URL's, folders and filenames actually have little to no affect in Google anymore. They are bolded in the SERP's for display purposes that hopefully the site is about what you have searched, highlighting some parts of the results for you to make your choice a little easier without having to read each and every result word to word.

Again, this is Google. Yahoo and MSN are completely different again. This is from actual testing I have done, most likely something like David has also done for himself to find the fact from the fiction.

Next part, hyphens. Google will actually read words within words, whether hyphenated or not. Google has done so for a long time now actually, for those that haven't noticed. You DO NOT need to hyphenate words within domains for Google to have them read as unique words. For example, the domain "thedomainisfortestpurposesonly.com", Google will actually read any part of any word within that domain. Google will see that exactly the same as "the-domain-is-for-test-purposes-only.com". To a user, the hyphenated version is easier to the eye, though to the search engine, either is acceptable and read the same for the same query.

[bobmutch]

Anthony Parsons:
"To begin with, an allinurl query has nothing to do with rankings. Searchers do not use an allinurl when searching. You might, but 99.99% of searchers do not." To say the allinurl: command has nothing to do with rankings would be the same as says the allintitle: and allinanchor: commands have nothing to do with Rankings.

Personally I don't use any of these in searches unless of course I am searching for sites to trade reciprocal links with. I have not put forth that people use, or should use the allinurl command in searches. I think most people don't even know what it is.

I hold that there is Ranking weight in having your keyword(s) in the URL. My understanding of the use of the allintitle: allinurl: and allinanchor: commands is to show you via the results which has the highest Ranking weight for a keyword or key phrase. If we take allinanchor:keyword it will show via the results, which page Google deems as having the highest Ranking weigh for keyword in the anchor. In my opinion the allintitle: and allinurl: commands work the same way.

My point was when we look at the results of allinurl:keyword that we will find that the results that the command produces are sites where the keyword is not part of a compound work but is a directory, a file or dashed out of a file or directory.

"Bob, I would also like to see your evidence. I have tested this theory and found the exact opposite." Evidence that Google gives higher weight to words dashed in the URL for what ever Ranking weight there is for having your keyword in the URL is shown in the 2 following examples.

Example allinurl:web
In the first 10 pages results you will find it doesn't once take web from webdesign or website but it is aways a dashed off another word or a directory, a directory of its own, or a file name. (In the case where there is one of the above it will also highlight the work in a compound word also.)

Example allinurl:web design
Same thing, in the first 10 page results you will find it doesn't once take web design from a filename or domain named webdesign but it's aways a dashed off another word or a directory of its own or a file name. (In the case there is one of the above it will also highlight the work in a compound word also.)

Right or work I consider allinurl: to be used the same way allintitle: and allinanchor: when it comes to giving Ranking weight for keyword(s) in the anchor text, the title and the URL.

"The bolding of terms is nothing more than a representative that the word was found. Google DOES NOT use the description for ranking purposes, though will bold the words found within the description if used instead of a snippet." I am not discussing bolding in a common search. I am discussing bolding in an allinurl: search, and more than the bolding I am discussing the Ranking of results in that search and the words in the URL of that search which is what Ranking weight for keyword in the URL is all about.

"As David said earlier, names in URL's, folders and filenames actually have little to no affect in Google anymore." I would agree totally. In fact besides the title all other on-page factors have little Ranking weight in my opinion. Little when compared to inbound links that have the keyword in there anchor text. And to open another can of worms, inbound links that have the keyword in there anchor text that are from relevant pages.

"Next part, hyphens. Google will actually read words within words, whether hyphenated or not. Google has done so for a long time now actually, for those that haven't noticed." I am aware of this and have noticed it for some time also.

"You DO NOT need to hyphenate words within domains for Google to have them read as unique words." I agree with this also. A simple search shows this. But where we at this point disagree, till I am shown different, is that when it comes to the Ranking weight given to keyword(s) in the URL, the Ranking algo deals with keywords in the URL the same way the allinurl: command does. It see how that is, just go and allinurl:web design and you will see just how that is.

If you can prove to me that the Ranking algo doesn't give weight to the keywords in the URL the same way that the allinurl: command does I will recant.

[bobmutch]

Mel:
"Ok Bob so to get back to the point what is your evidence that the bolding of words in the allinurl: search reveals the way that google is ranking for those keywords? " It really doesn't. It is the results that the allinurl: gives us and the urls in those results.

"My question why and what evidence can you show us that bolded words are anything but a display convention?" I have always maintained that the results positions from an allinulr: search and the keywords in the URLs of that same search tell us the story, bolding is only a display convention. It wouldn't matter to me if their were bolded or not.

[Anthony Parsons]

Ok, let me show you a good example, first one I picked, and even against my own site.

allintitle:internet directory

My site, http://www.anthonyparsons.com (http://www.anthonyparsons.com/) PR6, is currently positioned #7.
Yahoo, http://dir.yahoo.com/Computers_and_Internet/ PR8, is positioned #11.

You being such a PageRank nut and all, tell me who has the most authoritive page to rank higher for that term? Me or Yahoo? I reckon Yahoo, but they don't in an allintitle query. They do in a normal query for "Internet Directory". Why?

--------------------------------------------------------------------------

allinanchor:internet directory

First things first. Lets look at #29 with the URL terms "internet directory" highlighted as such www.einternetdirectory.co.uk (http://www.einternetdirectory.co.uk), no spaces, no hypens, but highlighted for an allinanchor search as I stated previously. Because your example/s didn't find any, my first one did, proving what I have earlier stated.

Now, lets do some third party checking shall we? PR Weaver Advanced Backlink Check for the first 100 results only.

My site comes up #17 for an allinanchor:internet directory.

# Has 1330 backlinks that are shown by Google
# Has 63,900 backlinks reported by Yahoo (just for statisical overall purposes)
# Has 27 unique backlinks that contain "Internet Directory" and I haven't counted all the other variations that where shown, such as "web directory" "anthonyparsons.com Internet Directory" "Internet Web Directory" etc etc etc.

The site above me at #16 for an allinanchor:internet directory ( http://jewishstudies.virtualave.net/index.shtml ) -

# Has 71 backlinks that are shown by Google
# Has 149 backlinks reported by Yahoo (just for statistical overall purposes)
# Has 2 unique backlinks that contain "Academic Jewish Studies Internet Directory"

My point! How exactly did that site get above me for an allinanchor check as a more authoritive site when it has less actual anchored links with that term, less backlinks in general, has a PR4 compared to my PR6, representative of the backlinks to the site? I could keep going, but I think I have proven my point on this one also.

I'm not even going to bother with the other, as I have better things to do than prove your theories incorrect. My point is Bob, is that your theories are wrong and you need to take our advice on this one, as already stated. I think you have found what your looking for, but didn't check thoroughly enough to actually deem whether the test had negative affects also. Its always easy to find a positive when you disregard any negatives, or just don't go looking for them. The queries do not work in the method you have described, clearly from the quick evidence I have shown above.

[bobmutch]

Anthony Parsons: "tell me who has the most authoritive page to rank higher for that term? Me or Yahoo? I reckon Yahoo, but they don't in an allintitle query. They do in a normal query for "Internet Directory". Why?" You Rank higher with your page on an allintitle:internet directory search as Google considers that your site has more keyword in the title weight toward Ranking than the Yahoo page does. No mystery there.

"allinanchor:internet directory... ...Because your example/s didn't find any, my first one did, proving what I have earlier stated." This proves nothing. I was referring completely to allinurl: searches. You are useing allinanchor:. You need to go back and read what I posted again. Try doing an allinurl:internet directory and you will not find one occurance of internetdirectory in the URL in the first 100 entries. (I didn't look any further that the first 100.)

"allinanchor:internet directory... ...My site comes up #17 for an allinanchor:internet directory... ...The site above me at #16 for an allinanchor:internet directory ( http*//jewishstudies.virtualave*net/index.shtml... ...My point! How exactly did that site get above me for an allinanchor check as a more authoritive site when it has less actual anchored links with that term, less backlinks in general, has a PR4 compared to my PR6, representative of the backlinks to the site?" Very simple Google considers jewishstudies.virtualave*net/index.shtml to have more inbound links with "internet directory" that you do. jewishstudies has 52, you say you have 27 so I will take you word for that with out checking.

I am not clear on the point you are even trying to bring out here Anthony. I post on the allinurl: command and you come up with a number of allinanchor: command searches and you don't even seem to know how that command works. If some page has a higher position in the returns for allinanchor:keyword it is because Google deems they have more or better inbound links with the keyword in them. That is cut and dry. It is the same with allintitle:keyword . The page that has the highs position in an allintitle:keyword search it is because Google considers that page to be better optimized for that keyword IN THE TITLE.

Now back to the subject at hand. The allinurl: command is the same. Google returns the page that has the best "Ranking weight for keyword in the URL" first. Man this is ABC stuff. You been up for 36 hours again : )

Now the question that is up for discuss as I see it is this. The allinurl:keyword command returns dashed and directoried keywords not combined or compound words or parts there of. Anyone that does an allinurl: search can see that plain as day. The question becomes, is the results of the allinurl: command indicitive of how Google handles the "keyword in the URL ranking weight"? I say yes. Do I have any proof. Yes! Simple - the allinurl commmand works the same way the allintitle and allinanchor commands do.

If that is so (and I am open to be corrected) then internet-directory.html will get more "keyword in the URL ranking weight" than internetdirectory.html will. Having said this though I will note that I don't think the Ranking weight it gets matters awhole lot. Links rule! Links, links and more links. Links with the keyword in the anchor text and on pages that are relevant.

[bobmutch]

pageoneresults: Sounds like a balanced view to me. Me I am not balanced : ) I am overboard.

My URL is going to be...

carsinlondon.com/forsale-london-ontario.cfm/ID/182/for-sale-used-cars/car/chevrolet/monte-carlo

I was going to put the year in but I desided to not do it. Don't want to make the URL tooooooooo long : )

[Anthony Parsons]

Ok, I thought I may have proven my point with the first two that you where playing with, though I will also with the last.

allinurl:search engine

http://www.google.com.au/search?q=allinurl:search+engine&hl=en&lr=&start=40&sa=N

Position #47
www.searchengineworkshops.com/articles/search-engine-seminars.html (http://www.searchengineworkshops.com/articles/search-engine-seminars.html) - 9k -

What your talking about is wrong Bob.


Man this is ABC stuff. You been up for 36 hours again : )

You obviously need to go back and learn your ABC's Bob, as my point is now shown to you, clearly. What you are saying is not how it works at all, and you are providing incorrect information to users. I think the above example demonstrates that enough.

Your reasons for the first two are also incorrect. Where did you get 52? The check I did didn't come up with 52, only 2.


Anthony Parsons: "tell me who has the most authoritive page to rank higher for that term? Me or Yahoo? I reckon Yahoo, but they don't in an allintitle query. They do in a normal query for "Internet Directory". Why?" You Rank higher with your page on an allintitle:internet directory search as Google considers that your site has more keyword in the title weight toward Ranking than the Yahoo page does. No mystery there.

Ok Bob, then what abour Position #47?

Yahoo! Directory: Computers and Internet > Internet (http://dir.yahoo.com/Computers_and_Internet/Internet/)
dir.yahoo.com/Computers_and_Internet/Internet/

Technically speaking, this PR8 page that includes the terms as much as mine, should then rank over me for an allintitle as it is technically more authoritive than mine.

Now if PageRank doesn't matter, and its all about the title, then this one back at position #80 should then also technically rank higher than myself or Yahoo as the phrase is mentioned twice:

Internet Directory - WeDoHosting.com - Directory Of Internet ... (http://isp-directory.wedohosting.com/)
isp-directory.wedohosting.com/ - 27k -

What I am saying is this. The order that Google displays an allintitle, allinurl, allinanchor has nothing to do with importance at all really. Because when you statistically analyse some of the sites, importance as we know it goes out the door, with higher backlinked pages above and below lower one's, the words are mentioned more within titles and URL's, etc etc etc. Your ABC's are not what you think Bob.

Maybe this is really something for the Google Guy to shead some light upon, because nothing that your saying is actually factually proven with everything we know that provides a page importance!

[Anthony Parsons]

Planning a site's navigational structure, I am wondering about folders and terms. Does MSN (& other search engines) rank a site differently if a key term being searched on is repeated within the entire URL.

EX:

blueshirts.com
bludeshirts.com/discount-shirts.html
or
blueshirts.com/mens/discount-shirts.html
And to boot, I'm not really sure what the hell allintitle, allinurl or allinanchor that you have raised has anything to do with the original question or the answers following the original question up to yours.

David answered the question. NO! It does not make enough difference for ranking purposes. It does make a difference for click through purposes from some A B testing I have done for myself, but that's about it. Unless your really chasing a highly competitive term, then its not worth the bother of a long and ambiguous URL string.

Yahoo and MSN do actually use it to a little more effect than Google does, but still not enough to warrant making long query stings for terms that you can capture simply by including the phrase within a heading, page and one anchor link pointing to the page. Yahoo and MSN are just that easy to manipulate. Hell, you don't even need the one anchor link really, its just nice.

[Mel]

So far in this long and convoluted discussion I have seen nothing exept supposition to support the idea that the allinanchor: allinurl: allintitle: or allintext: searches show us anything more than just what Google says they do:

Google restricts results to pages containing all query terms you specify in the anchor text on links to the page.
(you can find similar quotes for the allintext, allintitle, allinurl searches)

These search operators simply select a group of pages with a certain common criteria ie the allinanchor: search returns only pages that have the search term in links pointing to the page. I see no evidence whatever that they are then ranked soley in order of the importance of the allinanchor terms to the exclusion of all other ranking factors.

Without a body of strong evidence to support the suppoition that they are then ranked soley by the selection factor this is misleading at best.

[Anthony Parsons]

Well said Mel. I totally agree. Bob has placed his supposed facts, we have shown him otherwise, thus meaning, there are no facts to begin with. This isn't about right or wrong, simply what is and is not misleading information. Facts are provable, fiction is not. There is nothing provable about what has been raised contrary to this topics actual discussion.

[bobmutch]

Anthony Parsons: "Ok, I thought I may have proven my point with the first two that you where playing with, though I will also with the last. allinurl:search engine...Position #47 www*searchengineworkshops*com/articles/search-engine-seminars.html - 9k -... ...as my point is now shown to you, clearly." If you take a look again, in #47 where you see searchengine you will also see search-engine also. But having said that you will from time to time find a link in an allinurl:key word where only "key" or "word" is in the URL or where you will find the 2 words in a combined word (keyword).

This how ever is not the norm and is the exception. Why does it give you the odd results with out the words in the URL - who knows. Why does it from time to time give you a entry with "keyword" (and with out "key-word" being somewhere else in the URL) - who knows. But if you look at the results over all you will see that for the greatest part (I would guess on the average about 98% you will not find it showing an URL where the 2 words (allinurl:key word) are NOT in a compound word.

"Where did you get 52?"
The results shows "jewishstudes.virtualave.net/ - 17k 3 Dec 2004 Cached - Similar pages" Go to linkdomain:jewishstudies.virtualave.net on Yahoo it shows 52. You may of used on Yahoo linkdomain:www*jewishstudies.virtualave*net which only shows 1 or you may have used on Yahoo link:http*//jewishstudies.virtualave*net which shows 3. The link command on Yahoo only shows the IBL for the page you specific. The linkdomain shows IBL for the whole domain.

"Ok Bob, then what abour Position #47?" As I noted above, #47 of http*//www.google.com*au/search?q=allinurl:search+engine&hl=en&lr=&start=40&sa=N has a search-engine in the URL also. www*searchengineworkshops.com/articles/search-engine-seminars.html - 9k - Cached - Similar pages. A better example would be #92 www*best-searchengine*com - 12k Cached - Similar pages ..

So that is 97% of the results in the first 100 are dashed or directoried, 1% has neither of the keywords in the URL and 1% has them as a compound word. Doesn't that tell you some thing Anthony? These exceptions are nothing new to me and don't change what the results show. These kind of results your are pointing out are around 2% on the average. I suggest you take harder look at the 98%.


"Technically speaking, this PR8 page that includes the terms as much as mine, should then rank over me for an allintitle as it is technically more authoritive than mine." I would disagree. My understanding of the allintitle (right or wrong) is that the allintitle: command is only looking at how pages rank for the "keyword in the title weight".

"The order that Google displays an allintitle, allinurl, allinanchor has nothing to do with importance at all really." The order the results are shown in those commands have EVERTHING to do with importance. If you Rank first in allinanchor:web design that means that Google considers your site to have the most or the most valued inbound links with the word "web design" in the anchor text. The same goes for the allinurl: and the allintitle: commands. I would consider this to be ABC stuff. I find it hard to believe that you can disagree with me on this.

Perhaps on this one Anthony we will have to agree to disagree. I have not mean any of my comments to deride you nor to ridicule you. I am sorry if they sounded that way. As far as going off topic I agree. This should be forked. It is my position that their is no problem going off topic on a thread if it goes that way naturally.

If this topic is important a mod can start a new thread, if it is not the mods can give a call for the posts to come back on topic. I would like to see input from others on the allinurl, allintitle, allinanchor commands. If I am completely out to lunch and off the deep end I will just regroup and change : )

My position is that allinurl:keyword, allintitle:keyword and allinanchor:keyword are commands that return results showing what pages in the index have the highest Ranking weight for keyword in the url, keyword in the title and keyword in the anchor text.

My over all point was that the allinurl: command results shows that URLs with the keyword(s) dashed, directoried or single worded rank the highest for "keyword Ranking weight in the URL", not words that are inside, on the end or at the start of other words with out a dash seperating them.

[Anthony Parsons]

Bob, I love a good debate, and that's all I see this as. Nothing more. I know you not here to ridecule, nor am I. So, lets get on with it.

Anthony Parsons: "Ok, I thought I may have proven my point with the first two that you where playing with, though I will also with the last. allinurl:search engine...Position #47 www*searchengineworkshops*com/articles/search-engine-seminars.html - 9k -... ...as my point is now shown to you, clearly." If you take a look again, in #47 where you see searchengine you will also see search-engine also. But having said that you will from time to time find a link in an allinurl:key word where only "key" or "word" is in the URL or where you will find the 2 words in a combined word (keyword).
But you originally said that it can't occur, or words to that affect. The fact is, is that people use hypens or folders more than putting all the words in a sentence, which is more along the commonsense lines why such terms are returned first. Nothing more, nothing less.

"Where did you get 52?"
The results shows "jewishstudes.virtualave.net/ - 17k 3 Dec 2004 Cached - Similar pages" Go to linkdomain:jewishstudies.virtualave.net on Yahoo it shows 52. You may of used on Yahoo linkdomain:www*jewishstudies.virtualave*net which only shows 1 or you may have used on Yahoo link:http*//jewishstudies.virtualave*net which shows 3. The link command on Yahoo only shows the IBL for the page you specific. The linkdomain shows IBL for the whole domain.
Now you have lost me. What does Yahoo have to do with the anchor text in Google? When you go and look at the anchor text pointing towards that site, you will only find the (2) instances of Internet Directory mentioned in Google, not 52 times. I could be wrong, but not from what my check returned. You can't use Yahoo results for Google mate.

So that is 97% of the results in the first 100 are dashed or directoried, 1% has neither of the keywords in the URL and 1% has them as a compound word. Doesn't that tell you some thing Anthony? These exceptions are nothing new to me and don't change what the results show. These kind of results your are pointing out are around 2% on the average. I suggest you take harder look at the 98%.
Bob, you are giving a half arsed theory, not actual facts. I have never actually heard another person come up with what you are saying, yet. So for something that you believe is the ABC of SEO, is not so ABC mate, as never before have I seen mentioned that the order of results from an operator query, returns the most authoritive sites for that operator query. Rubbish. Those queries are exactly that, non-relevant, non-rank related queries. Nothing more, nothing less. Ranking #1 for an allintitle has nothing to do with authority, nor allinurl, etc etc.

The authority for a site is based on a keyword phrase being returned in order of priority within a normal search, not an operator query. Thats a good one Bob. Funny, ha ha. If I did a search for allintitle:internet directory, and just plane old "internet directory, your telling me that the one that is returned in the allintitle search is the most authoritive. Titles have nothing to do with links and so forth, so that is a great example in this case.

"Technically speaking, this PR8 page that includes the terms as much as mine, should then rank over me for an allintitle as it is technically more authoritive than mine." I would disagree. My understanding of the allintitle (right or wrong) is that the allintitle: command is only looking at how pages rank for the "keyword in the title weight".
But if that where try Bob, then my example should of been ranked over me, as one of those example even uses the term twice in the title, thus giving it more weight based upon the title only! It doesn't occur that way, because what your saying is far far far from the facts. Operator queries are nothing more than a figured return (the number returned), not the order returned.

"The order that Google displays an allintitle, allinurl, allinanchor has nothing to do with importance at all really." The order the results are shown in those commands have EVERTHING to do with importance. If you Rank first in allinanchor:web design that means that Google considers your site to have the most or the most valued inbound links with the word "web design" in the anchor text. The same goes for the allinurl: and the allintitle: commands. I would consider this to be ABC stuff. I find it hard to believe that you can disagree with me on this.
I find it hard to believe that your using operator queries and authority in the same sentence.

My over all point was that the allinurl: command results shows that URLs with the keyword(s) dashed, directoried or single worded rank the highest for "keyword Ranking weight in the URL", not words that are inside, on the end or at the start of other words with out a dash seperating them.
I fail to see any relevant point made thus far, as you have shown nothing conclusive. You have given examples of what you believe, and I have given you examples of exactly the opposite to demonstrate your theories are not as correct as you might think.

An allintitle is a great example to use as nothing else should factor the equation. Why don't you test two pages then Bob, one with:

"internet directory internet directory internet directory internet directory" or some phrase repeated, and it should rank #1 for an allintitle search as it should be the most authoritive, being mentioned the most times. The only factors that can be measured in the title is:

# character / word quantity
# character / word position

Simple test. Why not use "search engine optimization" actually as the phrase? That is probably the best to ensure it has plenty of competition and so forth, and its not just going to be some dumb arse lame phrase with no competition. I bet it won't rank #1 for an allintitle search! I have stated the factors assessed above, and by that, it should rank #1 for an allintitle search by what your saying.

[Anthony Parsons]

Here's another one, a little closer to the homepage at position #16.

http://www.google.com.au/search?q=allinurl:gift+baskets&hl=en&lr=&start=10&sa=N

#16 gourmetgiftbaskets.com

Gift Baskets is a popular term to include together within a URL, thus just as I suspected, it would be easier to find URL's with the term together. This is not a low % based on what your saying, but more a low % based on many URL's nowadays just don't run the words in together as they don't exist as they are already taken, so people have to hyphenate the terms, thus making the larger percentage.

Again at #39 abcgiftsandbaskets.com

Again at #42 adorablegiftbaskets.com

Again at #44 designityourselfgiftbaskets.com

You also find the word "gift" & "baskets" taken out of many of the domains, not as a hyphenated word, but out of a continuous phrase. I think you need to reassess your facts mate, cause all the evidence is going against what your stating.

Please provide evidence! If you provide factual evidence, then nobody should be able to find to the contrary against it, as its factual, not fictional. I think you simply use a couple of phrases, found what you wanted to find and stopped looking at the more popular of the URL's.

[bobmutch]

Anthony: "But you originally said that it can't occur, or words to that affect. The fact is, is that people use hypens or folders more than putting all the words in a sentence, which is more along the commonsense lines why such terms are returned first. Nothing more, nothing less." I looked back and didn't see where I said it "can't occur". I did say that it didn't occur for one example in the first 100 but don't see where I said it can't. My point is that when you us allinurl:key word you get dashed, single worded and directored occurances the majority of the time in the same way if you use alinurl:keyword you will get keyword in non-dashed.

I think the best thing to do is get a few of the people in the know comment on what the returns from allintitle, allinanchor and allinurl mean. Do the results from allinanchor show the pages with the most weigh for keyword in the anchor text first or not. Until we can agree on that we are just going around in circles.

On my other points in my post let me take a closer look at that a bit later> Some of the logic doesn't seem to make sense even to me now. Late night posting : )

[Dave Hawley]

Cetainly no expert, but;
"showing what pages in the index have the highest Ranking weight for keyword in the url, keyword in the title and keyword in the anchor text"
Leave off the bit I bolded and I agree.

There is no way that Google is going to make it that easy for SEO's and Webmasters.

[Mel]

This has been discussed before and I have pointed you to the Google statement which says only that the allin... searches restrict the selection of the ranking pool to pages that conform to the type of search and nothing about how they are ranked.

Lets take as an example the search allintitle:search engine optimization

#1 ranking page title
Search Engine Optimization, Search Engine Placement

#2 ranking page title,
Search Engine Optimization | Free SEO Resource Site

#3 ranking page title
Google Search Engine Optimization Search Engines & SEO

#4 ranking page title
bruceclay.com - Search Engine Optimization - Tools and Free Advice on Ranking and Promotion

#5 ranking page title
High Rankings Search Engine Optimization Information and Search Engine Marketing Tips

#6 Search Engine Optimization • Web Site Submission • Search Engine Registration

and down at #9 we have this

Search Engine Optimization Tips

it seems clear to me that there have to be other ranking considerations besides the "ranking weight" of the page title that determine who places where in this search which returns an estimated 480, 000 results

[Anthony Parsons]

What Does The Results From allinanchor, allintitle & allinurl Commands Tell Us?


It provides a collective figure only (number), for how many sites are directly competing / utilising each particular term queried! Nothing more, nothing less.

If you go back to our previous examples about allinanchor Bob, when you mentioned the site above mine has 52 anchored links saying "Internet Directory" and that I stated I had 27, approx, saying the same thing from the first 100 results only. The site that ranked above me had 52 anchors of a total 75 +/- links total. My site has about 250+ of that anchor for the total 1100 + links being read. How does the first rank over mine if I have more of the backlinks? Authority, this site has more than this site, etc etc has nothing to do with anything.

You have read to far into it. It is only a collective number displayed that is the desired outcome, nothing about ranking. Mel has shown you the exerpt from Google themselves, I have shown you examples of one ranking over the other when mathematically, they should not off, if you are only equating the one factor.

Google uses god knows what to place those in the particular order they do, not just what the query ran was for. If it didn't, then sites would run from order of characters vs position for all of those queries.

URL's would be ranked for how many times the phrase appeared + in what position it appeared.

Titles would be ranked for how many times the phrase appeared + in what position it appeared.

Backlink anchors would be ranked in how many times the phrase appeared + in what order of the anchor it appeared.

Etc etc etc etc. It was to easy to find immediate examples where each one of the above is not the case, thus the determination you have derived is not factual, only fictional by your understanding.

[powerofeyes]

There is no way that Google is going to make it that easy for SEO's and Webmasters. Agreed,

All those searches simply implies those sites are relevant to those advanced queries, I see far important ranking criteria than those factors

[Dave Hawley]

newreality, the question I asked myself when considering to name page as per their content was, why wouldn't I? The only answer against was it can look spammy if too many words.

I tend to believe it does help in the SERP's, but more importantly is the fact that searchers tend very much to scan the SERP's for kewyords they have used. When/if they spot them in the URL it can encouarge them to click the link.

Keyword pages and URL's are one of those things that users cannot cloak easily, that is, they are very much WYSIWYG. This is one of the reasons that I believe Google will give some weight to them. How much? Not sure but I would say not all that much.

Also, it is very helpful from Webmaster organization perspective.

In a nutshell, as long as is kept to < 3-4 words, I say go for it!

[Anthony Parsons]

That's why I tacked in that little disclaimer Bob, "or words to that affect", as I didn't really want to go back and read it all again, just took an approximate from memory. Disclaimers are wonderful things! :)

[bobmutch]

Anthony: LOL. Ya but you have to make the disclaimer a bit more broad cus I didn't say words to that effect either. I have searched enough time with allinurl and messed around with it to know that a small percent of the returns are not dashed, directoried or single worded.

I am going to see what kind of comments I got over on the thread it started before I go any further just to double check and make sure I am not completely out to lunch. I am pretty sure I know a few SEO heavy weights that hold the same position so I am checking with them.

Anthony you use McDar's tool much. http://www.mcdar.net/KeywordTool/keywordtool.asp Check it out and do a few searches with it. It shows the top ten for a keyword, with Google BL, PR, Site Pages, Allinanchor & Yahoo BL. It is very interesting how close allinanchor Ranking order follows the Ranking order for a search term. Check the tool out if you havn't. SEO-Guy has one almost the very same also. http://www.seo-guy.com/seo-tools/ultimate-seo-tool.php
I am checking with both of them and a few others on my position.

[Marcia]

If we use Google's advanced search and look for the exact phrase with allinurl:

http://www.google.com.au/search?q=allinurl:++%22gift+baskets%22&num=50&hl=en&lr=&output=search

we can see that they will pick up the exact phrase only where it's hyphenated and not run together. We can also see that it's exact without using stemming because the singular isn't bolded. Nor is it when capitalized in the first letter of the words, it's only bolded all lower case and when the words are separated by hyphens.

That doesn't tell us whether or not it's used for ranking purposes, or conversely whether using repetitions could hurt.

[Anthony Parsons]

I don't generally use free tools for two very good reasons:

#1 - There about as sharp as a Baboons backside, &
#2 - There about as accurate as a 2 year old boy trying to stand up peeing into a toilet bowl (not very).

Next your going to tell me you use Alexa ratings to determine accurate traffic statistics!

Maybe its just me, but honestly, you just have to look at all the free tools available, use each one and most of the time you end up with a different result. How is that possible if they are accurate. Then go and compare those results manually and you get another different answer. I don't find them much chop at all mate.

[dannysullivan]

As noted at the top of this thread, I've merged relevant discussions elsewhere to this one. My compliments, by the way, for everyone having a debate and discussion yet not getting personal.

[bobmutch]

Anthony Parsons: LOL, tell you to use Alexa. Na that won't tell you anything about your site other than if nerdy geeks that have alexa tb on are visiting your site. I get 400 visits a day and I think my alexa ranking is 28k which is a joke. It means nothing.

Well hey if have the time to go and check by hand the Google BL, Yahoo BL, PR, Pages, Allinanchor for the top ten results for a keyword fill your boots. Next thing we will hear is you are going house and buggy : )

Me I like being able to find these things out in like 1 min vss 15 mins. But like you say you are retired 33 and got lots of free time on your hands.

[AussieWebmaster]

The bolding is just what Google does to highlight the search term in the results. What the various contributions are are not directly realted to the bolding... though in some cases it is factored (such as title, etc)...

I think the searches bob suggests do contribute to insight and at varying levels contribute to the algorythm but to what extent etc. that is the mystery.

[bobmutch]

AussieWebmaster: What is your opinion on allinanchor:keyword . I have noticed when I do searches with the McDar tool http://www.mcdar.net/KeywordTool/keywordtool.asp that quite often the allinanchor Ranking is very close to the Google Ranking for keywords.

Keyword: Download
GOogle ranking:1,2,3,4,5,6,7,8,9,10
Anchor ranking:1,2,3,4,5,6,7,8,9,10

Not all are that close but with the keyword in the anchor text having such high ranking weight in the 10 and 10 of search I have done and completed the Ranking with the allinanchor ranking they are always very close.

I have heard a good number of other SEO people say that allinanchor shows how is tops for having the keyword in your anchor text.

[AussieWebmaster]

Well I think it reflects the query you are testing for but does not in all of those cases add to what will be the top result.

Do a search for currency trading... the one that gets top for allintitle, allinanchor, but nowhere for allinurl is second for straight search... and the one that is one for the straight search also does not get placed for inurl...

[bobmutch]

AussieWebmaster: "Do a search for currency trading... the one that gets top for allintitle, allinanchor, but nowhere for allinurl is second for straight search... and the one that is one for the straight search also does not get placed for inurl..."

No supprise there. The keyword in the url and in the title don't have anywhere the Ranking weight that the keyword in the anchor text has. That is why we see pages that don't have the keyword in the title, URL, or copy first page in the normal results just based on the keyword being in the anchor text.

#2 for computer http://64.233.161.104/search?q=cache:W8wjTkP4L48J:www.dell.com/+computer&hl=en is an example of this. I though it was interesting that it was #2 for allinanchor . This tells me that being #2 in the allinanchor means this site has the 2nd most weight for that keyword in the anchor text in the index and that is why it is #2 in the normal search.

[AussieWebmaster]

AussieWebmaster: "Do a search for currency trading... the one that gets top for allintitle, allinanchor, but nowhere for allinurl is second for straight search... and the one that is one for the straight search also does not get placed for inurl..."

No supprise there. The keyword in the url and in the title don't have anywhere the Ranking weight that the keyword in the anchor text has. That is why we see pages that don't have the keyword in the title, URL, or copy first page in the normal results just based on the keyword being in the anchor text.

#2 for computer http://64.233.161.104/search?q=cache:W8wjTkP4L48J:www.dell.com/+computer&hl=en is an example of this. I though it was interesting that it was #2 for allinanchor . This tells me that being #2 in the allinanchor means this site has the 2nd most weight for that keyword in the anchor text in the index and that is why it is #2 in the normal search.


But for currency trading the number one for anchor is number 2 in search... behind the site that is number 2 for anchor...

[bobmutch]

AussieWebmaster: "But for currency trading the number one for anchor is number 2 in search... behind the site that is number 2 for anchor..." Well that goes to show it is not totally keyword in the anchor text. Oh get this LOL. The number #1 has PR8 and the #2 has a PR6. There you go. See PR can still made the difference.

[Mel]

IMO the "ranking weight" for the page title is at least equal to that of a single instance of anchor text.

Anchor text ends up dominating because there can only be one instance (legally) of the page title but there can be an unlimited number of instances of anchor text pointing to the page.

[Anthony Parsons]

IMO the "ranking weight" for the page title is at least equal to that of a single instance of anchor text.
I would most likely disagree with that actually. Yes, you can use anchor text an unlimited time, though you have to look at it equally, on a 1 to 1 basis, and not quality vs. quantity.

The page title carries a lot of weight in its instance compared to anchor text, as anchor text is variable upon what page you put it on, how many links are already coming from that page to how much weight the engines place upon that link / page, etc etc etc.

On a comparitable scale, I would say the page title actually carries more weight than anchor text.

[Mel]

I would have to disagree that the weight of anchor text varies depending on the page it comes from etc.
IMO anchor text weight is unaffected by the page it comes from how many other links that page has etc etc.

Google seem to think that they are pretty much the same , putting data from both in the short barrels and considering page titles as links when telling you that the words searched for on a cached page are found only in links pointing to that page.

I see no evidence that there is any sort of themeing or relevancy calculations going on with reference to anchor text.

[Dave Hawley]

IMO anchor text weight is unaffected by the page it comes from how many other links that page has etc etc I find it hard to believe that the 500+ anchor text links to my Home page from my site would be given the same weight as 500 anchor text links to my Home page from 500 different domains.

[Anthony Parsons]

I'm not debating about themeing of pages and so forth, more that anchor weight is affected by many characteristics, for example, IP, Qty from IP, Qty links per page, position of link upon page, and other factors that have been tested and made public within forums. An anchor text link from a PR10 site compared to the anchor text link from a PR5 site will actually provide a different ranking comparison to the page they are pointing at. The title is a constant, which compared to an anchor would actually weight differently upon the page it is placed. All things being equal, the title would actually provide more weight in most cases than an anchor.

Don't get me wrong, anchors a very weighty, but on a one to one comparison, I would have to definately say that the title carries more weight.

You can prove this with a simple test, which has been done many times.

Add a page to your site, link to it from your homepage with just the URL as the anchor, which won't provide any anchor text attributable to the overall ranking, with just the keyword in the page Title. Record result once steady.

Now, remove the link from your homepage, and if what your saying about all anchors being equal, then simply place an anchor text link on some external page from your site and then measure the ranking result again. Remove that link once done, and place the same anchored link from another site and record. If the anchor weight was given the same, then regardless of what page you put that anchor upon, the same ranking would be achieved.

You will find this is not the case though. You could even make the test more accurate, by setting up two unique domains, both linked to from the one page. Place a link from one with the anchor to our page in question. Record. Now remove that one, and place the anchor on the other page, with ten other links to different pages above it and record. I bet you get a noticable difference in result. You could then even place our anchor text link at the top of that page, with the other links under it, and I once again bet you will get another different result, as link position is entered into the link analysis equation.

If you do this test for yourself, you will find the multiple variances that are derived from such link and anchor weighting in unique and varied circumstances. Give it a go, and you will retract the thinking that all anchors text is weighted the same.

[Dave Hawley]

IMO, such a test would be very inconclusive. There is too much natural movement in Google's SERP to do as you suggest and come to any definite conclusion.

While I do agree that the anchor text is not constant from all/any page. I don't believe the number of links and/or position would matter. IMO, the PR could well matter, but more importantly self-serving anchor text from GuestBooks, Forum Sigs etc would be given little/no weight as would 500 anchor text links from the inner pages of a site the anchor text points to.

[Mel]

Interesting Idea Anthony but there is no way you can seperate the various factors out that make up the ranking of the page to determine which factor is affecting the ranking.

It may well be that gaining an anchor link from a PR8 page will improve your ranking more than the same anchor link from a PR6 page but the differences could be put down to the PR component affecting the ranking independantly of the anchor text, and even at that not all PR8 pages have the same PR.

I cannot conceive how the IP is affecting the weight of Anchor text or how many links are coming from that IP, as this would totally destroy the rankings of sites hosted with virtual hosting, which is clearly not the case.

[Marcia]

This thread needs a bump right around this time, to take another look at how linking patterns are affecting Google rankings.