View Full Version : Dumb Google theory I thought of.
06-18-2004, 02:18 PM
Ok, heres a dumb theory I was thinking of.
Please do not laugh, I am sensitive. :mad:
From what I have read, Google places a type of point system on your page, scoring points for titles, keywords, links, etc....
I was reading about keyword density on a page, and from what I read, there really is no particular set density that google likes. Some rank really well with 7% keyword density, and others rank really well with 15% density. I have seen an auto site ranking high with 25% density. So from what I see there really is no set "Perfect" Keyword density.
But I figure there is still a way it has to play into things, so heres my stupid theory.
If we know a page can have absolutley nothing related to what ever the "keyword text" in the link was, but be ranked #1 with enough votes for that keyword (for example the little "miserable failure"/george bush thing"),
Than is it possible that If I build a page with enough keyword links to make me #1, (say the keyword is wall clocks) and My content does have the same keyword (wall clocks) as the link did, Google will check my keyword density ratio and determine the ratio I have , and set it as being the Ideal ratio to score against any other page concerning wall clocks??
Then if someone submits a diffrent page , about the same keyword, google will score the sites keyword value , using my ratio as the the "Ideal Page"
Just some more rantings from me, but thank you for reading.
06-19-2004, 05:50 AM
What you have stated is basically the same theory I have believed for a while. I think Google determines the "ideals" for things like KW density and all other factors based upon pages that it determines are the best for a particular search term. All other pages are then scored based upon those parameters. The only place I would differ from your theory is that Google probably uses more than jus the #1 site, and it probably also uses historical data, so that the longer a page had been ranking well, the more it would influence what the "ideal" is.
Of course, the above is all just my own speculation, so everyone can take it for what it's worth.
06-19-2004, 07:41 AM
To me this theory seems to say-the #1 site is the best because its the best. A site would be ranked #1 based on certain parameters...then you are suggesting that after google ranks the site #1 it then looks at that site and uses the defining characteristics of that site to rank all other sites?
To me that doesn't make any logical sense. :eyes: If google has certain characteristics to rank the #1 site in the first place why would they then need to use some arbitrary attributes from that site to rank all others. Why not just rank them using the same characteristics they used to determine the #1 site deserved to be #1. It would seem to me to be a waste of server resources.
Plus, what about sites that have been #1 for 6 months or 1 year...or more, but then suddenly fall. Why would that be?
There have been many markets I have gone into and I have been able to "unseat" the #1 spot-and with this theory how would you explain that? Did I match the ideals of the former #1 guy...but then have more links? To me that seems quite coincidental.
I think a better theory is that maybe google is using how long people stay on a site to determine ranking. We know that google has the ability to monitor this-via the toolbar-and it would not be a resource hog to determine it...in fact it would be less of a server strain than PR currently is. And what is more appropriate to determine relevance than how long a searcher stays on a given site. All google has to track is after they click on a result how long is it until they are back at the search page. Sites that users find interesting (relevant) remain in the listings and sites that users quickly click away from slowly disappear-along with all the other google data (PR, links, anchor text, H tags, kw density, title tags, etc.)
Just my thoughts,
06-19-2004, 08:33 AM
I'm thinking more of an iterative process, rather than a simple single-shot one that you are describing. It's actually not that different from the local rank process we know Google uses from their description of US Patent 6,526,440 (http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=1&f=G&l=50&co1=AND&d=ptxt&s1=6,526,440.WKU.&OS=PN/6,526,440&RS=PN/6,526,440), where the results actually given to the user are based upon a preliminary group of results that are then re-ranked based upon (in the case of the patent) interconnectivity. Google calls this "refining" the results. It could just as easily be done for any other factor.
It is possible that Google uses some data from the toolbar, the problem is that it does not provide a random sample of users. Since it is only available on one browswer for one operating system, the results could easily be skewed. Think about pages ranking for linux terms -- the pro-windows, anti-linux pages would have way more toolbar users than the pro-linux pages.
As far as bumping out the top-rated site, I have no doubt that backlinks remain king, and these other factors are weighted much less (but still weighted).
Agian, though, I stress this all just opinion and speculation (but without those we'd never be able to discuss anything about Google).
06-19-2004, 09:51 AM
Oh, I agree...re speculation. I enjoy this type of discussion. It is fun to bounce around ideas.
I didn't think of it in terms of local rank-hmm...interesting. I am going to have to reread the patent app. I will freely admit that I haven't read any google whitepapers or "treastise" in probably 4 months. It may be about time to reread some of these papers.
Oh, what fun that will be. :)
Actually, this theory doesn't look dumb to me at all. For a while I've always thought that Google changes their ranking criteria depending on which keyword you are targeting, they won't rank "wall clocks" sites with the same criteria as they would with "sports betting" sites, I'm happy to see that someone actually mentioned this, now I know I'm not the only one and I don't feel dumb, hehe.
It does make logical sense that they would rank pages differently, you can't use a pre-set ranking criteria for all pages since all pages sell different products within their targeted area of course. :eyes:
06-19-2004, 10:59 AM
[QUOTE=pagej]I enjoy this type of discussion. It is fun to bounce around ideas.
I agree -- that's what's so great about this forum. The only way we can figure things out is to hear different opinions.
06-19-2004, 04:21 PM
Thanks for the positive thoughts on my theory, its really nice to see.
The thought that this theory may be possible is based on the simple thoughts similar to an if+then statement.
Now keep in mind that I only assume that the google system rates on keywords, a point score, and then adds up all your point scores for your page/site to determine rank.
But in my theory, all I am stating is that I think that if google finds a large # of "votes" towards a site via links, that what ever the text was in that link must be important on that site. And if all these people are linking to it, it must be a good site for this "word" so google will check the percentage of keywords on this page, and determine that any other site refrencing this keyword, should have a keyword% similar to this site, in order to score maximum keyword points.
Of course no matter how much is scored in keywords, does not matter, as the final page score is the most important.
I've been comenting this whole thing with a coworker of mine, he's an SEO as well, according to what we been talking, is that linking towards a page, and also, the algorithms used by Google to rank each targeted keyword changes. For example, in my field, sports betting, the algorithms are very complex, due to the fact that its a highly competitive field, but for something that's very unique, the algorithms will be simpler to comply with. What I wander is, if you follow the top pages on their format, it'll be hard to outperform them, so there's got to be a way to not follow them but be higher, now isn't that the answer we're all looking for, duh! hehehe.
Anyways, its great to be part of the forum, I'm sure we can all find that way.
BTW Ride, I see Eibach Springs as ur signature, u a car freak like me?
06-22-2004, 10:37 PM
Car freak, Yes sir I am!
Thats my business, car suspension.
Hope you get a chance to visit my site wnhen done.
Right on! Well, just lemme know when its done, I might just become a customer of yours
vBulletin v3.0.3, Copyright ©2000-2013, Jelsoft Enterprises Ltd.