Content Quality??

I wonder how G calculate quality of content?

I can think of how G can differenciate unique content. But quality means how resourceful that content is. How G find out worth of words?

 

Search engines cannot judge the quality of content. Sometimes even humans can't do it. You can read 2 articles let's say on health by two experts, the articles will push totally different views on some issue, both articles will cite research supporting their views etc. How do you judge who's right? You can't unless you dig out the research, see the study design, know statistics well etc. etc. Search engines simply cannot judge the quality based simply on words.

Search engines can:
1. Detect dup content.
2. Detect unnatural wordings like overstuffed with keywords text.
3. Judge the quality of the incoming links
4. Track how much time visitors spend on a given page/site.
5. Track on which pages do searchers click when searching.

Basically, search engines try to judge the quality based on links and user behavior.

 

I wouldnt be so quick at saying they cant dudge good content.
With the amount of money going into semantic searching and stuff, Im pretty sure google has some good ways of judging what is good content and what isnt...
It will never be as good as a professional, but there must be ways to estimate...

 

I agree.

With the money and minds they have invested in their technologies, they must have at least some basic ways to judge the quality of content. Links are another good indicator, putting exchanged or paid links aside. Obviously if someone links to some content, it must have value, either good or bad, but value.

 

With Latent Semantic Indexing (LSI) scanning everything I wouldn't be surprised. Also, I know Google does have actual people looking at sites and "Grading/Categorizing" them. The Job has been listed on their site numerous times.

 

They will certainly be getting closer to this fact.

I cant remember exactly where I have seen it but there is a massive AI nearing initial completion that is pretty complex.

They have trained it to know things that we would take for granted as common sense. Things like if a glass has water in it and you tip it upside down, then it will loose or spill the water.

They have been loading it will information from top specialists all around the world. It also has the ability to learn from data it is given. Also there will be a free online interface and a scripting language you can use to make it do things.

Very impressive.

Incidently because they had loaded alot of information about history and past people. It came to the conclusion that everyone before 1890 was ritch and famous.

lol Obviously not 100% but just shows that we are moving in that direction.

If I find the link I will post it.

Brad

 

They can certainly expand queries, detect naturally written text against artificially generated/keyword inflated. BUT, natural writing does not guarantee quality, as I may naturally write the dumbest content in the world. We are very far away from having computers understand text in the way we do. That's why the best way imo is to concentrate on user behavior. We tend to stay more at quality sites, browse them, read them, link to them, click on results leading to them etc.

LSI can discover relationships between words and their meaning, but has nothing to do with the quality of information. LSI can be used to expand queries so that when you let's say search for "calorie counter", engines can look also at "calorie counting", "count calories" etc. related terms. LSI can be used to put more value in natural writing, but again, natural writing does not guarantee quality and truthfulness of the information presented.

 

I think their primary methods of grading pages for quality is spelling and grammar. There might well be more to it than that but that would go a long way as I see it.

Add to that an analysis of word counts, relative and related words (keyword density and presence of related words) you'll end up with a simple yet effective quality score.

 

Essentially, basic spam detecting techniques have a fairly accurate grade on the quality of content.

Any obvious key word / phase stuffing, and other spammy things make for poor content for spiders and for people. Poor spelling could relate to poor quality of writing.

So many basic things can eliminate content from what could be considered good. Obviously there are exceptions, but something like this would definitely be a good basis for grading.

 

I understand that; I ma just saying if they can create an algorithm like LSI what will they create next?

 

they don't know

 

it judges smart, when google staff is banning a site.

 

They are moving towards a more automated process in which the spiders and algorithms will allow search engines to judge the quality of content. Will it ever work out? Probably not in the near future. PR was supposed to do this to some extent, and they will probably continue to tweak how links treat the SERPs

 

can you tell me what is the google defination for quality content??

 

I have both kinds of sites. Ones that I deem to have "quality content" and ones that are pure crap. They both do well. I had heard that google has state of the art content filters that nothing gets past. So I decided to test their filters myself. My lord what got through was ridiculous. It seems to me

"He was very happy."

to google reads just as well as...

"My shoe was spectacularly elated with anger."

Doesn't make sense but will Google catch it? My results say no.

 

guys, they have to process huge amounts of info extremly fast, to jump to some sort AI system is way in the future