I did a search on LSI and see that nobody is talking about it here. Latent Semantic Indexing is what Google used to knock all those AdWords Spammers on their butts recently.

Is anyone here using LSI to help improve their SE rankings?

Originally Posted by steubing

I did a search on LSI and see that nobody is talking about it here. Latent Semantic Indexing is what Google used to knock all those AdWords Spammers on their butts recently.

Is anyone here using LSI to help improve their SE rankings?

Yes. I generally target 3-4 key words or terms on each page that has enough copy to warrant this. The issue, however, is how on-point Google is with determining "related" terms. In my experience, they're very hit-or miss.

I read on SEO2020 of Charles Heflin that you can do that:

(~) with your search term, and Google will highlight allthe related terms.

You must add all these them in your site with a proper Silo structure.

I did that, and it's working, because Google see those words as related terms.

It is called theme coverage.

If you are so great on LSI, do you know that LSA (from which LSI derived) also has drawbacks and *limitations* of use? No wonder, nothing 100% perfect under the sun.

And if it has resrtictions of use, then LSI should have restrictions in application.

My 2cents.

Originally Posted by freetraff

If you are so great on LSI...

Who here is claiming to be "so great on LSI"?

No technology is perfect when it first comes out. The recent Google Adwords "Slap" is proof that LSI is being used and is going to get used more and more in the future. If perfected, Google and the other SE's will be able to us technology like LSI to prevent web site spammers from getting into their index. Those who know how to build their websites to please LSI and the SE's in general will be handsomely rewarded with good SERPs.

steubing, my bow to LSA when it comes in full force, before that - it has tweaks. For example, you have a 100% dup text, you add to it simple RSS newsblocks - and no problems, no longer you are dup, I thought for LSI it'll be easy to smash this - NO.

And Google slap. To tell the truth, I saw a lot about it on the forums, but did not feel it. I have a bunch of testing sites that are complete *nothing* in terms of good content or design, I use them to see what works and what not, and good content can be an extra factor for the clear test. Nothing happened. Surely I don't use AdSense on them, maybe that was the factor, but all the rest like indexing and top positions preserved.

And after the horrible Google slap I saw one site with primitive cloaker having top position for a competitive keyword. And nothing. The big G was snoring.

So, summing it up. I agree with you that LSA should have big future, because the idea is more than good, but A LOT has to be done. From what I see now - this will be a long road. Before it happens Google will still need to hire students and pay $15-20 per hour for manual refining of search results.

Charles Heflin is selling a book on LSI http://www.seo2020.com

Someone bought it?

I follow his forum, when I am free, and the guy is good.