Forum Moderators: open
1) when i do a search on my site with google using the "similar pages"-command in the toolbar i get only 30 guestbookresults and 3-4 similar results! this shows me that the main theme for my website is not my most relevant keyword, but guestbooks. so when you sign guestbooks (no matter if you use your keywords in the link text) google will interpret it in this way that your site is about guestbooks! (why would i otherwise get guestbook-results when using the similar pages-command!)
2) my keywords NEVER appear in the top ten of the serp's although all results above me have lower pr! you don't believe it? sign guestbooks and you'll see! it's really bad, because my keywords are everywhere on my pages (title, meta, alt, filenames, url, link anchor, keyword density, just everything! i used brett ranking tipps)
i would do everything to get a pr4 instead of my pr6 if i could just remove this stupid links from the guestbooks - it was the biggest mistake i could ever make and i really advice everyone who is playing with guestbooks not to do it.
i better get a new domain - all my work since august has been destroyed just because i was so stupid to sign guestbooks :(
Did you say an Angelfire site with 5,180 Google-credited inbound links and a PageRank of only 2? :o
What search was this for?
It isn't that non-competitive. Yet, they are still page 1, and beating out over 5,100 sites for it.
5,100 sites returned for a single keyword is extremely uncompetitive. Pick any 10 words out of the dictionary at random and on average you're likely to find at least that many returns. (unless the word is absurdly obscure)
Looks like this page got there by having this keyword as the only word on the title of the page, plus it is in the H1 tag big and bold at the top of the page.
It's common knowledge that the title of a page is one of the single largest factors in the Google algo and has been for ages. [webmasterworld.com]
Low competition + Undiluted Exact Keyword Match in Title = Piece of cake (with the addition of a bold header it's even more of a shoe in)
How many other sites in the top 100 have that single undiluted keyword as the title of their page?
Based on what you've described I don't see anything surprising at all about their landing squarely on the first page.
im curious as to what people here believe is 'competitive". my main keyword, botox, is 177,000. With britney spears at 2,390,000, and porn at 72,400,000, and my name at 209 (lol)... what sort of competition are -you- going against?
If it is much lower than that, not many people will be searching for it, and beyond 1.5 million my ranking is not good enough to get much traffic.
But you also have to remember that just because a search returns a low number of results, doesn't mean that there aren't many searches for it (1.e. a newly released product). But in most cases it will be fairly accurate.
oddly enough, I do not get alot of traffic from it.
I get most of my traffic from a #1 with 19,500 returned pages.
By doing this, you'll dilute the "theme" that google seems to have given your site due to these guest book links while strengthening the actual theme or your site.
While it would take time, in the longer term, you would receive the PR boost from these guestbook links while still maintaining a strong theme about your site due to getting on-theme reciprocal links.
Just my two cents worth if you don't want to kill off your domain or attempt to get rid of all your guest book links.
Jim
Since then, people have figured-out ways to try to manipulate off-site criteria as well. But the fact remains, that it is still a lot harder to manipulate external sites than it is to change your own HTML.
A few comments:
1) Google indexes pages, not sites, so a Webmaster can manipulate PageRank to favor certain pages. If you were to have a site devoted to recipes, for example, you could very easily set up your site's structure so that affiliate pages for pots and pans, kitchen appliances, ingredients, and cookbooks would get a bigger share of PageRank than noncommercial recipe pages.
2) I'm not an SEO expert, but I get the impression that some Webmasters are pretty successful at manipulating "off-site factors" through the use of multiple domains, carefully planned crosslinking, multiple listings under different business names in DMOZ and Yahoo, etc.
3) I suspect that Google is capable of doing a much better job of discerning on-page spam (keyword stuffing, for example) than search engines were just a few years ago. Natural-language analysis has many applications such as OCR, machine translation, database queries, and helping to determine authorship of literary texts. I'm sure you can add Google search to that list of applications.
Check you stickymail. That is the number of pages this search term appears on.
When searching for the keyword phrase you sitckied me, I see your index page at #10 and another page of yours at #11.
Search Google for "allintitle:kw phrase" and the sites listed above you have shorter titles, you may be diluting those keywords in your title with the length.
Search Google for "allinanchor:kw phrase" and it will show you the sites having this kw phrase in the anchor text pointing to them.
I think the above may have more of an influence on your position than the fact your site is listed in some guestbooks. Just my thoughts.
You can also start to "dilute" the guestbook links by getting more on topic links from similar websites, niche directories, etc.
I agree with chiyo's comments, and I think Google builds the "similar pages" based on the content (titles, anchor text, description) found around your incoming links, so whatever is near your link in those guestbooks you signed will probably show up in your similar pages. It may be that the more often some of these other sites are listed in close proximity to yours, the more relevant they become when calculating similar pages. This will start to change as you get more relevant links. I know, I've been through this. It can change. I always look at where my link will be placed before requesting a link. Especially on those pages that are only arranged alphabetically. Where would you fit in?
Go to Google WebQuotes [labs.google.com] and search www.yoursite.com. Not sure if this will become relevant to rankings or not, but it is interesting:
"Google WebQuotes annotates the results of your Google search with comments from other websites. This offers a convenient way to get a third party's opinion about each of the returns for your search, providing you with more information about that site's credibility and reputation."
Yeah, but this is beating out over 5,000 sites, some that in fact *are* well optimized for that keyword, and have decent PR. Sure, with a term that only matches 5,100 pages if you try hard you should be able to get very high on the SERPs. However, this page obviously *didn't* try, and is doing well against some pages that did, and have good PR.
>How many other sites in the top 100 have that single undiluted keyword as the title of their page?
I think you may have hit on it here. The Google algo likes that it is *undiluted* more than the fact of the page's PR. The algo must weight this highly. However, odd that it is beating higher PR pages that have that word in the anchor text linking to them in many more cases. The only page that links to this with the word in the anchor text is the sole internal page that links to this page.
[edited by: rfgdxm1 at 12:08 am (utc) on Jan. 6, 2003]
A lot of this would depend on not how many pages this term appears on, but also *are those pages trying to score high in SEs on that search term*. Here's one I just picked at random:
[google.com...]
Results 1 - 100 of about 201,000,000.
Huge number of pages with that. However, if some SEO registered index.cc (this is available), and aggressively fought to do well on just that one term, they might succeed. The reason there aren't many pages fighting to be #1 for just "index". However, best of luck getting to #1 on Google for "computers". ;)
Some interesting reading:
What is this page known for? Computing Web Page Reputations
Paper by Alberto Mendelzon at the University of Toronto:
[cs.toronto.edu...]
HTML version in Google cache [216.239.39.100]
More references:
[citeseer.nj.nec.com...]
[citeseer.nj.nec.com...]
It's common knowledge that the title of a page is one of the single largest factors in the Google algo and has been for ages.
That certainly seems to be true now, but was not always the case. A year ago, you could get Top-10 rankings on a COMPLETELY UNOPTIMIZED page for searches returning 20-50,000 results, just by getting a PR5!
deft_spyder, "what is competitive"...
#2 4.5 million :)
Oddly enough, I don't seem to get much traffic from this one either, mat_bastian.
I would say anything under 10,000 is non-existent, 20-30,000 is pretty easy. 100,000 and up starts to get competitive, and things start to get serious around 1/2 million. Trying to get #1 for anything past 1-2 million starts getting into the stratosphere, where luck and "which way the switches are flipped" at Google can have a significant impact. A lot also has to do with how competitive the field is. Getting #1 out of 500,000 would be a lot harder for a field that would be more likely to do SEO (such as financial, computers, etc.) than for a low-tech area (such as cow pie tossing)
1) Google indexes pages, not sites, so a Webmaster can manipulate PageRank to favor certain pages...
This is true, europeforvisitors, but you are only shifting-around what you already have. The total PageRank for a site is still dependent on links from external sources.
Off-site manipulation is still possible, just a lot harder. The "tricks" you mentioned are considered spamming by Google and/or DMOZ, and can get your sites perm banned. On-site is much more subject to manipulation, and all they need is a few minutes with a text editor!
RE: stickymail -
Sorry all, I had been gone from this site for several months, and I forgot all about the stickymail feature! I just found out I got a bunch of msgs, give me a chance to read 'em and I'll get back to you all.
Googleguy hinted some time back that one of the recent tweaks in the algo was to de-emphasize PR to combat the practice of PR for sale. This page is an example of the devaluation of the importance of PR.
>I would say anything under 10,000 is non-existent, 20-30,000 is pretty easy. 100,000 and up starts to get competitive, and things start to get serious around 1/2 million.
The case I mentioned of that Angelfire page involves a keyword that is the trademark of a pharmaceutical. Nothing so competitive that any half way competent SEO shouldn't be able to blast his way into the top 10 with some effort and decent PR. But a totally unoptimized page on a free host with negligible PR?
I therefore these days see PR as value only in the power of that page to pass on page rank to pages linked from that page.
It is only sensible that pages returned by a query are ranked by relevance primarily, though Pr from "linked to" pages does give some boost. But at the stage of actually perfoming a query, the PR of the pages you actually see in the SERP are almost irrelevant by that time.
That to me has always explained why lower PR pages seem to be above others in many SERPS. PR measures the credibility of a page as an authority and how much PR is can pass on, irrelevant from themes. Ranking in SERPs indexes how relevant each page is to your specific query. Sometimes, but not always, an "authority page" can also be a relevant page!
Which may explain why HHufRuFru pages are not being ranked well in SERPS.
Is that correct?
The issue is precisely how much importance should PR be as part of the algo? Google rather recently has definitely turned down the dial that controls how important PR is in determining SERPs.
PageRank lost much or even most of its usefulness when Webmasters learned about it and started optimizing for it. One could argue that the Web and Google would be better off if Messrs. Page and Brin had kept their Ph.D. research under lock and key and and Google had never incorporated the PageRank meter into its toolbar.
I don't see how Google could go back to giving significant weight to PageRank without excluding whole categories of sites from PR calculations.
Come to think of it, maybe Google should ignore inbound links from .com sites while giving a big PageRank boost to .edu domains. Financially strapped universities could build new libraries, laboratories, and football stadiums by auctioning their links to commerce sites on eBay. :-)
But what's more interesting about HuHuFrufru's topic, in spite of the difficulty of the situation, is looking at WHY it's been decided that his site is about guestbooks.
Does the link text pointing to his site say "Guestbook?" Probably not. So what makes the guestbook pages where those links reside seem relevant for "guestbook," which is apparently accruing to his site? Is it density on the guestbook pages? Probably not, they're generally not optimized as such and the word is most likely not prominent. It isn't other guestbook pages linking to that one.
It's probably in the page title and possibly in the filename for the guestbook pages - and more than likely appears in link text pointing to the guestbook itself, possibly from high PR pages in the site. There could be an if/then condition that counts a link more valuable from a high PR page without altering the nature of PR itself.
So if HuFru's being considered to have a site theme of guestbooks, then that's being determined by the context of the linking pages. No, we're not seeing themed or context-sensitive Page Rank yet, though some are predicting it, but it doesn't have to be that. It just has to be the right alchemy among the factors involved.
Thinking about it, three of the things that come up a lot in talking about Google's scoring are page titles, keywords in URLs (and domain names), and keywords used in link text. Thinking about it further, if those factors exist to any degree on the linking out pages, should that exclusively define what the linked_to page is about if the particular keyword or themes related phrase doesn't appear on the linked_to page itself in some of the significant page elements?
The biggest weaknesses I see with Google are the relative ease of manipulating Page Rank, and far easier and more common, blatant manipulation of link text.
It seems that putting more weight on congruence between theme-related factors appearing on both the linking and recipient pages could dampen the effect of the factors that are individually skewing search results we see, particularly in the heavily SEO'd areas we're always hearing about.
To draw an analogy, it's almost like a site's good reputation getting ruined by a lot of nasty "gossip" being spread around town in theh form of malicious links. Someone can have their "reputation" ruined without a shred of guilt on their part.
This kind of thing almost makes it look like an experimental tweak to see how bad it can get.
This is one problem for search engines. SEs want to produce the most relevant results for users, and SEOs the most relevant results for clients. Once webmasters know to optimize for PR, then to increase relevance Google has to increase the weighting of factors that they don't optimize for. I also think a lot of this has to do with the dominance that Google has attained as a SE rather recently. As Google has ascended in dominance, do has optimizing for high PR.
The issue is that since the beginning of search engines, at some point some people started to discover that by doing certain things they could get higher rankings than other sites. Some are just good at it, regardless of which seach engine is the object of greatest focus at any given time.
The whole SEO industry came into being with the emergence of search engines long before and irrespective of PR, and the same phenomena would exist without Larry Page or Sergey Brin. It did before and it always will.
The existence of factions is an incontrovertible fact of life: the search engines themselves, those who know how to optimize for good rankings, those who don't like those whose optimizing beats them in the SERPs nor the search engines they're being beaten at, and the public, who make the final decision by using the most relevant search engine, and most importantly, by pulling out their credit cards and exercising their purchasing power.
Whatever's commercially viable is what will win out. There's a cat and mouse game that'll go on forever no matter who protests, and there will always be some mice who end up getting more of the cheese than others. We can more or less like it or lump it, but that's life. ;)
Thanks for the 5 paragraphs Marica, but that one sentance did it all;)
But what's more interesting about HuHuFrufru's topic, in spite of the difficulty of the situation, is looking at WHY it's been decided that his site is about guestbooks.
What is "Similar page" links
[webmasterworld.com...]
and on signing guestbooks
[webmasterworld.com...]
It's common knowledge that the title of a page is one of the single largest factors in the Google algo and has been for ages.That certainly seems to be true now, but was not always the case. A year ago, you could get Top-10 rankings on a COMPLETELY UNOPTIMIZED page for searches returning 20-50,000 results, just by getting a PR5!
Completely unoptimized is fine... It's not Google's intention to reward "optimized" sites, it's their goal to reward what they regard as the most relevant pages. Period.
PR ascribes reputation... other factors ascribe relevance.
PR alone could never ever cause a site to rank without at least showing up on Google's relevance radar screen. Otherwise we would only get PR10 returns in the SERPs regardless of the keywords searched for.
As for title weight, the following post will clearly show that despite your claim, it most certainly was one of the single largest factors even a year ago:
Brett's post from April 19th 2001:
[webmasterworld.com...]
As for those suggesting that the "similar pages" results have no bearing on Google's algo... think again.
Much like the variety of command strings they offer (link: allinurl: allinanchor: etc) you can bet that Google's not wasting their precious bandwith and resources maintaining and displaying results from indexes which they don't consider important. The question isn't whether it's important or not... it's simply a question of how important.
If they didn't feel that the similar pages results were relevant in most cases, you can bet they would not be featuring the option in their SERPs.
Of course there will always be some exceptions. That comes with the territory. Their goal is always going to be serving most of the people, most of the time. Complaining about the exceptions isn't constructive, and doesn't mean they are failing in their goal.
That said, a thoughtful study of the exceptions can help us to understand their methodology, and in turn help us to work with it... rather than against it.
I have looked at his site and I agree with him, there is an extra factor that has gone into his rankings. This is a new factor, becuase he states that his rankings were FINE prior to the lastest google update. What we have to try and discover here, is what this new factor is!
I believe it is theming of content of pages from which a page receives links. I have looked at the sites that are ranked above his, and the only thing that makes sense to me is Google's increase in weightage of theming of inward link pages.
I suggest that some of you also analyse this closely so that we can discover what this EXTRA factor is.
do you understand? the only difference is the guestbook-thing. that's why i said it must be the guestbooks!
i don't know, rfgdxm1 suggested to me that it could have been my index.htm which has nearly no content (the real content starts at the 2nd page). but why does the other site (which has also a no-real-content-index-html) so well?
(why no content on index.htm? my site is non-commercial and i wanted to have a beautiful start-page. in august when i started webdesign i didn't know that optimizing for search engines is sometimes more important than for humans)
[edited by: mat_bastian at 2:29 pm (utc) on Jan. 6, 2003]
