33 replies to this topic

[BBCoach]

I've found "related" has changed quite a bit, in the past results were based on links so where linkage was "promiscuous" the results had no discernable relation to the sites' "theme", today they seem a lot more focused.

It goes back to the early days and "Google Scout", when the guys still believed links would invariably be on-topic and so "related"

Agreed, that's what I've been seeing over the last few years and there has been an improvement (with G in particular) of getting a more accurate grouping of sites.

It was also before there was AdSense, AdWords, and Universal Search which requires them to know the theme/category of a web site (and before link spamming became such a problem). Also, throw in a powerful data gathering tool GoAn. I think we can all agree that G already groups web sites by location and that is a category/theme in and of its self. Also, if G has implemented the 2005 patent "Traffic prediction for web sites," then they're "tagging" or categorizing web sites as Ian alluded to previously.

A good question is how much of an impact, if any, does it have on rankings? At this point I can't make a determination, but I'm not ignoring it either. Why? Because based on groupings by geographic location my domains render dramatically different results when searching at the different locations and I want to determine how deep the rabbit hole goes and it's impact on rankings. I could be wasting my time and it wouldn't be the first time.

[projectphp]

It goes back to the early days and "Google Scout", when the guys still believed links would invariably be on-topic and so "related" :-)

That is a false argument in which a too high standard is applied. 100% ontopic and 100% related is neither the goal nor a requirement for using links in an algorithm. After all, if it was, a link from a car store to a car stereo store would not meet the 100% requirement, and that sound prtty bloody useful a link to me!

Links STILL, even with ten gadzillion SEO sites, offer invaluable information vital to the most practically effective algorithms (as oppossed to the mythically effective like having people review every page ever second). Noise, rather than a deal breaker, is simply a problem to be factored in.

It was also before there was AdSense, AdWords, and Universal Search which requires them to know the theme/category of a web site

No, it doesn't. AdSense and AdWords don't require a category at all, and Universal Search only requires that a site be tagged as belonging to a specific subset of sites, just like video, blog, linux, government, many of which have been around for donkey's. That there is a blog search DOES NOT mean that categories are used, needed or even preferable in other areas (like say cars), and especially not that "theming" is in use.

(and before link spamming became such a problem).

Is it a problem? And if so, as "before" implies a start date, when did the problem start? Making a statement like that is mioving to fast, You need to quantify it. besides, ranking websites is really just one big problem, with hundreds of sub-problems, of which link spam is just one minor issue. Saying somethign is a problem then is either a truism (i.e. trivially true) or a statement that something is harmful without evidence.

A good question is how much of an impact, if any, does it have on rankings?

Does what have? Hehehe.

At this point I can't make a determination, but I'm not ignoring it either. Why? Because based on groupings by geographic location my domains render dramatically different results when searching at the different locations and I want to determine how deep the rabbit hole goes and it's impact on rankings. I could be wasting my time and it wouldn't be the first time.

I am going to say something controversial here: nothing that any SE has done has changed SEO one bit since about 1997. Nothing. The game always was and always will be:
1. Build a Search Friendly site with consistent URLs.
2. Write naturally using words people use when talking about the thing you are writing about (aka stop speaking Marketingspeak).
3. Get links to your site from others.

That is the sum total of the game. Sure, there are lots of ways to maximise SEO, but the above is what it all boils down to, and if you take the long term SEO view, you will already be doing most of what is required to get the best results. I say that because The SEs are trying to get the best sits to rank the best, so if you aim to be the best, you are what they want to promote.

It is kinda like trying to become the right person for a lover. You either are right for someone or you ain't, and while faking it can get you quite far for a while, ultimately the truth wins out.

[Jill]

I am going to say something controversial here: nothing that any SE has done has changed SEO one bit since about 1997. Nothing. The game always was and always will be:
1. Build a Search Friendly site with consistent URLs.
2. Write naturally using words people use when talking about the thing you are writing about (aka stop speaking Marketingspeak).
3. Get links to your site from others.

Yep, definitely agree here. And it's been since before '97 even I think.

[glengara]

"That is a false argument in which a too high standard is applied."

Interesting, you'd wonder why they chose a linkage based tool to find similar content if there was no such expectancy....

[BBCoach]

No, it doesn't. AdSense and AdWords don't require a category at all

Then I guess it's magic on how Google does this.
"AdSense for content automatically crawls the content of your pages and delivers ads (you can choose both text or image ads) that are relevant to your audience and your site content—ads so well-matched, in fact, that your readers will actually find them useful." I know they simply pick a couple of keywords on your site and match it up that way. Yeah that explains it.

(and before link spamming became such a problem)

I should have qualified it with paid link spamming. Link spamming started back in the later half of the 90s, but was soon made moot by the algos and human reviewers by 2001. For the depth of the paid link spamming problem you're gonna have to ask Matt Cutts how prevalent it is.

[projectphp]

Interesting, you'd wonder why they chose a linkage based tool to find similar content if there was no such expectancy....

Expectancy of what? Either I don;t get your comment, or you don't get mine. Can you clarify what you meant?

Then I guess it's magic on how Google does this.

So does that mean you agree with me that they do not using theming or categories? Or is everything a theme, thus rendering the word useless?

I should have qualified it with paid link spamming. Link spamming started back in the later half of the 90s, but was soon made moot by the algos and human reviewers by 2001. For the depth of the paid link spamming problem you're gonna have to ask Matt Cutts how prevalent it is.

Well, no, that didn't clarify at all! You seem to now be saying link spamming isn;t a problem, because and I quote "Link spamming ...was soon made moot by the algos and human reviewers by 2001". So which is it, cause I am really confused!

[projectphp]

Yep, definitely agree here. And it's been since before '97 even I think.

Hehehe, one has to be conservative in estimates occassionally Jill!

[BBCoach]

Well, no, that didn't clarify at all! You seem to now be saying link spamming isn;t a problem, because and I quote "Link spamming ...was soon made moot by the algos and human reviewers by 2001". So which is it, cause I am really confused!

projectphp you need to read in context. It might help your understanding. If you were involved back before 1997, then you would know about link farms and other link building tricks. If you were around in the early days, why play coy? BTW, are you saying paid linking schemes aren't an issue today for SEs?

So does that mean you agree with me that they do not using theming or categories? Or is everything a theme, thus rendering the word useless?

If you would have read what I earlier wrote "that it is not theming but categorization" you could get beyond that. Please read.

No I don't agree with you in total, but I don't have enough data to qualify either way and neither do you. That's why I'm running tests against it because it's not magic. I want to quantify the "how/why?" Can you explain how/why Google can do that?

And speaking of 1997, why did you pick that year where the SEs had done nothing to change SEO? If you're talking about the SEs base algos, then ok and it was much earlier than 1997, but to say that is not taking everything into consideration including improvements. It's easy to throw dates around, but quantify the statement please. Maybe we all could learn something about this magical year.

Edited by BBCoach, 25 January 2008 - 10:13 AM.

[projectphp]

So what exactly does your test entail?

[piskie]

That's a very very good question.
Since I read "That's why I'm running tests against it............" I have been wondering how such a test could be accomplished.

[BBCoach]

Why don't you answer my questions instead of asking a new one? Hmmm? I'm not going to tell you what or how I'm testing it. That's proprietarty information. The mere fact that I'm testing it should be enough to point someone in a direction. My questions still stand. Or are you full of quick and cute rebutalls? C'mon answer the questions. I might give some back based on your fully qualified answers, but if all you can do is ask questions, then I don't think so.

[projectphp]

clearly, you are upset. and seeing as we aren't really having a conversation, but what you seem to think is a "who can annoy who more" competition, rather than upset you anymore, I'll just say good night and good luck, and I hope your test is useful

[BBCoach]

No I don't get upset with people that I don't know, maybe frustrated but never upset. That is unless a stranger tries to kill me on the road by not paying attention while driving. Then you'll see me upset. I reckon my text could be construed as "upset," but in a conversation people go back and forth. You were clear in your stance that G does not categorize and I really do want to know "why" you believe that before I answer new questions. I want to know what you're basing your stances on versus just being "quick and cute" with your opposing views. Why? Because you may know something that I didn't think to test and therefore you would save me some time going in the wrong direction.

A little about my testing. I'm using two sites one an ecommerce site and one an informational site on two different servers. Both are setup with GoAn. Informational is setup with AdSense. Ecommerce site uses AdWords with PPC campaigns. Why use these two kinds of sites? Well ecommerce is where I'm most concerned so I had to use one in my testing model. Originally began testing with the ecommerce site. The informational site in my opinion would be the polar opposite of an ecommerce site and it has been completely reworked with its current version live for one year. I launched it with the purpose of conducting this test because I'm not able to use AdSense on the ecommerce site. Ecommerce site has several thousand link backs and is taxonomically built for the industry. Informational site has a few hundred link backs and is categorized by region. I've been testing (does G categorize and how deep) for almost three years making adjustments to each site to determine if it matters. Also both sites are 10+ years old to avoid any aging delay. BTW, I'm contemplating adding a third site to the testing mix if I can find the time to build it.

Your turn and I do apologize projectphp if my text offended or sounded like I was upset. Frustrated would be more accurate, but not really that much.

[piskie]

I hope you don't mind me butting in here BBCoach, but their apears to me to be a fundamental flaw in your described test.
I would suggest that an informational site describing "Widgets" is similarly themed to an Ecomerce site selling "Widgets"

IMHO, the Same Theme with Different Objectives and as such makes the test invalid.

[BBCoach]

Not even close piskie and I don't mind your input. They are two totally different domains with two totally different populations.