Semantic engines - nothing difficult, at all...

[freetraff]

When doing the test on one semantic search engine I was really shocked how easy it is to get good positions (ranks).

It turned out that:
- if you have article directory with relevant content
- sitemap and rss feeds submitted
- few backlinks from authority sites
...You are in the game!

So where the heck is this scary and frightening LSI and semantic wave that is going to clear the net?

Maybe some very-very primitive automatic sites that are nothing but pile of garbage text will be junked from the good positions. But there are almost out of the game in any case. And manipulating with content turned out even easier that I thought. Some content + On-Page SEO + at least some authority to the site = you are ninja.

I am frustrated and disappointed.

We need Artificial Intelligence. The talks and buzzes about 'beware' are nothing but talks and buzzes.

[Webnauts]

What about the documents semantical structure?

[freetraff]

Originally Posted by Webnauts

What about the documents semantical structure?

Hi Webnauts, can you please explain what you mean? What searches have I done to test?

If you clarify, I will be glad to explain.

[Webnauts]

Originally Posted by freetraff

If you clarify, I will be glad to explain.

Check my post in my forums.

[freetraff]

I think I understand what you are talking about ;)

I DO think that semantic is a new trend, and can help to properly "assess" the document, and correlate it with some niche. Linguistics is part of my education, so I really welcomed the blending of mathematics and linguistics in the new semantic trend as by that time linguistics had real proofs that it could be a very effective combination (don't think that all that semantic stuff popped up out of nothing, the combo of maths and linguistics has been working together over 50 years I think, just under another 'sauce').

So, thumbs up to search engines to look that way. Semantic approach can help to properly assess the document.

BUT it turned out so easy to fit their standards (read - 'get good positions in their ranks')!!

And this frustrated me a lot.

I was doing a tiny research on Hakia.com I will not go into all details, because if you are really interested you can read the whole search engine marketing report here on eZineArticles.

But if we are talking only about SEO part of the story, then my opinion that it's easier to manipulate this type of engines than Google. For example, from my testing experience I am 95% sure that if I build on my site article directory (as additional section of the site) around some niche with the tool that will be properly fitting page title, description, meta and content/article - semantic search engines will soon 'think' that I have very relevant and good content, that grows regularly, that will be fitting parameters of wise balancing of all elements on the page, etc.

Heck, article directory (or niche forum, or niche, nicely substructured blog), some time, and you are at the top.

To be frank, I was expecting a lot more. I know that semantic SEs give more sense, but the manipulations with sense are even easier than with the 'sense+link power' in Google (if I dare to skim its algo to this simple formula).

[kgun]

Originally Posted by freetraff

Linguistics is part of my education, so I really welcomed the blending of mathematics and linguistics in the new semantic trend ...

If that is your education, the links in your signature, as of may 7 2007, disappoint me.

"All Copies Sold Out!


But You Have A 'Limited Time' Chance To Get
New Updated Version Of
Auto Content Publisher For FREE!"

...............................................

"Auto Content Publisher is a unique alternative to pain and problems. It will automatically create content rich pages/articles for any keyword that is relevant to your web site topic. It will properly place the necessary links on created pages. It will 'tell' search engines that a new page/article has appeared on your web site and where search engines can find this new article to index it faster and include it into their listings. Auto Content Publisher will do this and much more in such a way that search engines will never delete your web site from their listings for search engine SPAMMING".

My bolding.

Questions:

1. Do you believe in this?
2. Where did you take you linguistics education?
3. Are you sure that you have fully understood LSI?

I think there is no free lunch, at least in the long run. Read this step-by-step-usability-guide first published by WebNauts in the "Web Site Accessibility and Usability" sub forum. He is fighting an almost lonely fight here about standards, and you are degrading the web by automatic content publishing.

[freetraff]

Thanks for the reply kgun and surely I would like to reply.

Originally Posted by kgun

Questions:

1. Do you believe in this?

I think you mean in article directories? Yes. They do really help. And they are a very relevant content medium if the directory is niche oriented. When all elements of SEO are together (onpage, content, RSS publicity of this content, regular growth, etc.) this is a content bomb.

About my article directory tool. I think it's not fair to boast about personal stuff and I will not. There is nothing ideal. But this tool automates much of the stuff, though I must admit the market went far beyond (and I released new tools, on some of the tool European investment companies build their business). What I can say is that I installed this tool over a year ago on one of my Forex sites and since that did not even play a lot with it. Site get's traffic and AdSense profits.

But it's important to say that article directory is great when it's like an 'add-on' to something else on the site. For example, on my Forex site I disclose one Forex trading system with all rules for free, and directry helps visitors to find more info. Google has over 800 pages indexed. And no problems (and this is without the duplication solution feature being activated).

But let's stop here. You had more interesting questions.

Originally Posted by kgun

1. Where did you take you linguistics education?

Kiev National Linguistic University (Ukraine), and as far as I remember from my education 'Quantitative Linguistics' emerged long ago - it's not that some folks started to do at the end of 1980s. And Quantitative Linguistics is from where the idea of blending math and language rooted.

This sphere of linguistics gave many brilliant inventions. For example, how to analyze the hidden psychology that stands behind the text and what author had in mind. Or how to tell if this text was written by male or female (by a special number of elements, later on this gave birth to gender linguistics).

What I am trying to explain is that I DO trust that this approach (LSI) has future, but...

Let's talk more about it in the following answer.

Originally Posted by kgun

1. Are you sure that you have fully understood LSI?

Nothing can be congnized completely, and if you see where I am wrong, please correct me. I will try to explain what I squeezed from LSI.

If we don't go beyond many tech details (I am not much of a tech guru). LSI, using the blending of linguistics (e.g. shades of meanings, rows of synonyms, clusters of relevant words) and math, is aimed at giving better understanding about 'what is this priece text about' (meaning and relevancy).

This can help a lot to search engines. Because it's not necessary that text with word 'fashion' is about fashion industry - it can be about Internet marketing like 'the new fashion of following up a customer with autoresponder proves...'. LSI is about solving problems like this, squeezing the real meaning of content piece.

Hence, when people search for something, LSI trended engines must give them better quality and relevancy of info.

And now the scary stories - "it will change the pattern of Internet marketing and SEO". LSI changes nothing. If you know how to play with content, make all its elements fit - you are the winner.

Article, good article, is written about some topic. Written by human being with brains, not some automatic junky text with keywords. LSI should adore articles, because article is a great fit for LSI - one piece of text should be playing about one key issue. Plus, articles are usually written by educated people, they will use many synonyms, and for LSI synonyms are a big helper in defining the topic of the content.

So, build a directory around one niche, do nice job about onpage optimization (like title in meta title, maybe title in the URL, proper meta description and meta keywords) plus add to this the real content on page. This is the best piece of meat for LSI.

Can it be article directory only? No. It can be a forum. Or blog where people post comments. But article directory is surely a great way to grow good content on the site and "give the meat to LSI processing".

And now let's take a hypothetical situation: site A with 5 pages of good content and site B with 100 pages of good content. If LSI is inclined to the content, who has more chanced to become a winner? Who has bigger potential to catch more surfers. Even if site A result is #1 in ranks, but then site B has 20 times more. And catches the surfers on all possible clusters of relevant meanings within this topic.

And it means that good old things still work. And no need to worry about big shifts in SEO and IM with the advent of semantic search strategies into the major SEs. And all scary talks and buzzes are just scary talks and buzzes.

P.S. By the way, from my experience of comparing Google and Hakia - I felt better about Google, because it was shortcutting me to the creators/sellers/vendors/providers faster - and many things in the net are about finding these sites. Yes, this experience has some psychological tricks, because maybe I got used to the Google way of searching. But I told what I felt and think that it's not something outlandish.

[kgun]

Have I met an Ukrainian mathematician, then I should weight my words :-)

After reading you excellent answer, your sig link surprises me even more.

Originally Posted by freetraff

I think you mean in article directories? Yes. They do really help. And they are a very relevant content medium if the directory is niche oriented. When all elements of SEO are together (onpage, content, RSS publicity of this content, regular growth, etc.) this is a content bomb.

Agree to that.

Originally Posted by freetraff

About my article directory tool. I think it's not fair to boast about personal stuff and I will not. There is nothing ideal. But this tool automates much of the stuff, though I must admit the market went far beyond (and I released new tools, on some of the tool European investment companies build their business). What I can say is that I installed this tool over a year ago on one of my Forex sites and since that did not even play a lot with it. Site get's traffic and AdSense profits.

Then we do not longer agree. There are two alternative tools that to my surprise are still living. I have deleted the links in my collection.

• the-article-system dot com
• articlebot dot com

This reminds me of Abraham Lincoln's words:
"You can fool some of the people all of the time, and all of the people some of the time, but you can not fool all of the people all of the time".

As an economist I would say that if you can earn money legally by fooling some people all the time, ...

It is no worse than that some young people without serious certification are declared genious in every bull market.

But I would personally expect more from a person with an education from Kiev National Linguistic University. Your links are much worse than your writing. I do not personally include them in my link farm.

You are who you link to and in a sense you are who links to you.

Originally Posted by freetraff

P.S. By the way, from my experience of comparing Google and Hakia - I felt better about Google, because it was shortcutting me to the creators/sellers/vendors/providers faster - and many things in the net are about finding these sites. Yes, this experience has some psychological tricks, because maybe I got used to the Google way of searching. But I told what I felt and think that it's not something outlandish.

I read that comparison. Interesting. You could have done even better, more thorough comparison, with your background.

Originally Posted by freetraff

This sphere of linguistics gave many brilliant inventions. For example, how to analyze the hidden psychology that stands behind the text and what author had in mind. Or how to tell if this text was written by male or female (by a special number of elements, later on this gave birth to gender linguistics).

Yes, this experience has some psychological tricks, because maybe I got used to the Google way of searching.

Is psychological bias a better word?

Or how to tell if this text was written by male or female

Like dating sites, where men with (manipulated) pictures contact hundreds of other men and hope to fool x per cent of them to send flowers to a flower shopper, pay for translation services, visa, travel, etc. etc. :-)

At least they should ask their sister to write the letter.

[freetraff]

Thanks for the feedback. It's a pleasure to hear sound and reasonable response.

Originally Posted by kgun

Have I met an Ukrainian mathematician, then I should weight my words :-)

To be frank, I also don't know many (not my major). But I know that some of the math guys here have been doing computer type machines (big and heavy like a truck, and worked 1 hour and then 1 week of repairing) in 1950s. This is not the earliest computer invention, but I think they did their best :) But big brains were deluded to Moscow (don't forget that it was USSR empire), that is why you might know some good Russian folks, but some of them can be Ukrainian. Though I am not much of an expert about math gurus; fairly admit it.

Originally Posted by kgun

This reminds me of Abraham Lincoln's words:
"You can fool some of the people all of the time, and all of the people some of the time, but you can not fool all of the people all of the time".

As an economist I would say that if you can earn money legally by fooling some people all the time, ...

It is no worse than that some young people without serious certification are declared genious in every bull market.

But I would personally expect more from a person with an education from Kiev National Linguistic University. Your links are much worse than your writing. I do not personally include them in my link farm.

You are who you link to and in a sense you are who links to you.

Sorry, kgun, lost your thought here. But what I can say, is that even that type of article tool was very ok for the market, I judge from the feedback. It was a bit tough in settings, but the results were really ok. Plus the forex site is also milking nice AdSense profits. This is true. I have my AdSense stats to prove that.

But when I say I lost your idea - I mean the following. I don't know what tool in the sig you are referring to, because there are two of them. And one of them is 1st page in Google (international) for 'link exchange script' and 'automatic link exchange' for several months. Not the toughest keywords on the planet, but this is the niche where everyone says to be guru for link building. And not so many get to the 1st page. And I think that it'll be impossible to get there with very bad links ;)

Originally Posted by kgun

I read that comparison. Interesting. You could have done even better, more thorough comparison, with your background.

Agree, there is no limit to perfection. But in the article I showed only part of the keywords. I used more in real test, just explaining all of them was a bit of bore. But I think I got the trend and semantic search engines should put more attention to the link power from authority sites. Otherwise Google will be the more useful. At least to me.

Originally Posted by kgun

Is psychological bias a better word?

Agree. But only this type of bias - psychogical. Because Google is not paying me money :) And if you check some of my posts, you can see that I like pointing to weak points of Google. And it has got plenty of them.

I was just really expecting a lot more from semantic engines. If we don't critisize their faults right now, we will never see better quality of search results. If right now they think that semantic approach will save them - they are in trap. I made some examples of how anyone can create good content pool for these engines. Then it means they are easier to be cracked than Google.

[Steven1976a]

Great postings

I especially liked this bit.

"And now the scary stories - "it will change the pattern of Internet marketing and SEO". LSI changes nothing. If you know how to play with content, make all its elements fit - you are the winner."

Surely SEO will never be one thing that makes you rank well. It surely needs to take into account everything to give the best results. Even 20 years from now i would imagine meta tags such as page title will pay a part although maybe only a small part along with keywords, good code, content, links etc....

[kgun]

Originally Posted by freetraff

Sorry, kgun, lost your thought here. But what I can say, is that even that type of article tool was very ok for the market, I judge from the feedback. It was a bit tough in settings, but the results were really ok. Plus the forex site is also milking nice AdSense profits. This is true. I have my AdSense stats to prove that.

Yes, there are lot of them. Try the following search on Google:

forex kgun site:www.webproworld.com

and you may see what I mean about them. Is your an exception?

Originally Posted by freetraff

But when I say I lost your idea - I mean the following. I don't know what tool in the sig you are referring to, because there are two of them. And one of them is 1st page in Google (international) for 'link exchange script' and 'automatic link exchange' for several months.

It was especially the second link about auto content publishing. But, automatic link exchange, does that still function? Then Google has still a long way to go.

But as long as you earn money legally, I understand you ...

Time to split the web in 2, one for robot generated content / articles, and one for content / articles produced by human beings?

[freetraff]

Originally Posted by kgun

Time to split the web in 2, one for robot generated content / articles, and one for content / articles produced by human beings?

And with the advent of artificial intelligence we'll lose the very hope to tell one from another :)

I saw some solutions that try to generate content automatically (via scrubmling paragraphs), but that looked very lame. However, I think with good AI everything is possible.

[kgun]

Originally Posted by freetraff

And with the advent of artificial intelligence we'll lose the very hope to tell one from another :)

Answer awaited.

Originally Posted by freetraff

I saw some solutions that try to generate content automatically (via scrubmling paragraphs), but that looked very lame. However, I think with good AI everything is possible.

No it is not, read the post with the following heading "Trust rank and trust" in my blog.

Then come back and give your comment.

P.S.
By the way, AI, is a subject where I have a little more than average knowledge, since I wrote a master thesis in mathematics about nonlinear, fractal and chaotic structure in financial time series (295 pages available in Norwegian only).

Make a page search (CTRL + F + Kjell Bleivik) on this page and you find a paper, "Determinism and stochastics in economics" that I presented there. Even if it is may 2007, I stand by my conclusion from the above blog post:

"Objective search is not an easy game. When you use mathematical algorithms, you must know what they say and don't say. You must know their limitations. That a lot of people vote for the same person, does not imply that that person is more trustworthy. It simply says that he got a lot of votes. You are not more right, even if most people agree with you".

A SE is excellent at indexing and finding content. A human being is still better at recognizing an airoplane from a bird, a cloud or a leaf. Why does a supercomputer not beat the best chess players?

I was afraid you was one of these :-)

[freetraff]

Originally Posted by kgun

A SE is excellent at indexing and finding content. A human being is still better at recognizing an airoplane from a bird, a cloud or a leaf.

That is why Google is not even objecting the facts that they use human beings to refine search engine results in some very competitive niches. (if I am not mistaken there has even been a post on this forum about this issue)

I thought about a way to have some kind of official "Watcher" from Google that is installed on each site and watches the behaviour of visitors. If they click a lot, and click 'wisely' (I know it's a vague term, but let's accept it for a while) - meaning do on page dowloading, read, go inside the site structure - then it's pretty clear that site is useful for the visitors.

But this idea has two drawbacks: (1) too many site patterns (click behavior on one page report and article directory is different) and (2) possibility to similate this process by robots (via some deep clicking into the site).

So, not that smart idea :(

I think Google tries another way now. Those who know about Google's new feature - Web History (have more about it on my blog) - which is a kind of 'gatherer' for the click behaviour. And from this we have a tiny step to checking if people who came from Google's search phrase stay on the site (if they stay - how and what they click) and if leave the site quickly - well, if too many people do that - maybe it's a reason for red flag.

Originally Posted by kgun

I was afraid you was one of these :-)

No :) Linguists adore having more fun. And these guys at least half of the day spend on the job. :)

[kgun]

Originally Posted by freetraff

I thought about a way to have some kind of official "Watcher" from Google that is installed on each site and watches the behaviour of visitors. If they click a lot, and click 'wisely' (I know it's a vague term, but let's accept it for a while) - meaning do on page dowloading, read, go inside the site structure - then it's pretty clear that site is useful for the visitors.

Official watcher from Google and Click wisely, yes diffuse concepts. Interesting theory though. If sites should have a official watcher, it should IMO, be a tool provided by W3C that can be used by SE companies.

Originally Posted by freetraff

But this idea has two drawbacks: (1) too many site patterns (click behavior on one page report and article directory is different) and (2) possibility to similate this process by robots (via some deep clicking into the site).

So, not that smart idea :(

Yes, I think it should be possible to simulate the process or behaviour. May be circle reasoning. The problem is technology, the software that the SE Bots uses. Then it is difficult to overcome this malfunction or whatever you like to call it by using technology. There are areas where humans are superior to Bots, and I think in the foreseable future.

Originally Posted by freetraff

No :) Linguists adore having more fun. And these guys at least half of the day spend on the job. :)

In a sense mathematics is objective, but is it this objectivity that give the results the user want in every case? Some times it is natural to use Google News and other times, The Wall Street Journal. Great that we have both. More meaningful possibilities and options, can not reduce the surfers webexperience.

What is a SE BOT without advanced mathematical alogorithms, and they are getting more advanced? May be we can agree that a relatively high dimensional, fairly advanced mathematical system, is a necessary, but not sufficient condition to produce quality results on the SERP's.

Since this is posted in the SEO forum, the message in your first posts is valuable and interesting. In a sense it confirms my simple SEO rule (a least common denominator). Make dynamic qualtity content with simple clean code and relevant foccused anchor text of IBL's.

Here is an interesting related article:

A New Search Engine Approach

[freetraff]

Originally Posted by kgun

Here is an interesting related article:

A New Search Engine Approach

Hm, copyright approach, interesting. Did not think about it this way.

Need to think it over, because I found tiny leaks in this approach, but maybe they are surpassable and/or superficial. Thanks for sharing this, Kjell, good food for the brain.

[freetraff]

Originally Posted by freetraff

Hm, copyright approach, interesting. Did not think about it this way.

Need to think it over, because I found tiny leaks in this approach, but maybe they are surpassable and/or superficial. Thanks for sharing this, Kjell, good food for the brain.

Kjell, this is what got brainstormed from the idea mentioned in the article (check my SEM ideas on eZineArticles.com). I think the levels of content authority can be of help and bring more logics to the market of website publishers.

[mothproof]

Originally Posted by freetraff

When doing the test on one semantic search engine I was really shocked how easy it is to get good positions (ranks).

It turned out that:
- if you have article directory with relevant content
- sitemap and rss feeds submitted
- few backlinks from authority sites
...You are in the game!

So where the heck is this scary and frightening LSI and semantic wave that is going to clear the net?

Maybe some very-very primitive automatic sites that are nothing but pile of garbage text will be junked from the good positions. But there are almost out of the game in any case. And manipulating with content turned out even easier that I thought. Some content + On-Page SEO + at least some authority to the site = you are ninja.

I am frustrated and disappointed.

We need Artificial Intelligence. The talks and buzzes about 'beware' are nothing but talks and buzzes.

speaking of sitemaps, i can't get mine working :(

[leaperm67]

I have recently constructed a suite of new UK finance industry websites. I have tried to employ what I understand is meant be LSI optimisation techniques, including siloing and theme's for each individual topic specific site. Can anyone point me in the direction of LSI testing software so that I can check if I have got it right and if not what needs changing.

Mark Leaper
www.moneymatchmaker.com

The sites are:
www.loans-moneymatchmaker.co.uk
www.mortgages-moneymatchmaker.co.uk
www.remortgages-moneymatchmaker.co.uk
www.endowment-compensation-moneymatchmaker.co.uk
www.insurance-moneymatchmaker.co.uk
www.credit-cards-moneymatchmaker.co.uk
www.shares-moneymatchmaker.co.uk
www.moneymatchmaker.com/
www.moneymatchmaker.com/loans.php
www.moneymatchmaker.com/mortgages.php
www.moneymatchmaker.com/remortgages.php
http://www.moneymatchmaker.com/endow...mpensation.php
www.moneymatchmaker.com/insurance.php
www.moneymatchmaker.com/credit-cards.php
www.moneymatchmaker.com/selling-shares.php
www.moneymatchmaker.com/uk-finance-news.php
www.moneymatchmaker.com/mortgage-guide.php

[kgun]

Here is a video by Google that I hope is on topic.