There is a lot of rumor that goes on in our industry that just isn’t fact, someone says something based on something they’ve heard, and then the game of telephone begins. Every time it is repeated it changes just slightly until it is way out of left field or just way misleading.
So I call those items the MYTHs of SEO, and someday I will list them all. But here is a big one. The Myth of LSI (Latent Semantic Indexing)
Not to toot my own proverbial horn, but while the LSI theory cooled the jets and put the fear of you know who into the ranks of would be seo’ers, I kept saying, “you know, I hear what LSI is, it makes sense, but guys, I am just not experiencing the effect… I get just as much of a link jolt from linking to a finance site from a rocking mechanic site as I do another themed queen bee finance site… wt*?”
NOTE: I will and do always speak from experience and my own testing OR I will tell you that it is something I have *heard* and give the source if I can dig it out of my sometimes dusty brain. 😉
So, I have been *experiencing* NO effects of LSI is in play.
And, this is now what I *hear* from seo Deity himself, Lesle Rohde, that Latent Semantic Indexing is Lame LOL In a recently released video, he shows his research on the subject (which is very in depth I might add) and gives a solid explanation why LSI just ain’t working. Here is the recap…
1) Contrary to popular belief, Google does not have one single Paten on LSI, or any that even reference it.
2) Patents that are commonly mistaken as LSI patents only reference phrase, word and occurrence techniques filed by one researcher in 2006.
verify at http://www.google.com/patent
3) And even though in 2006 there was a lot of talk about LSI, LSI has NOT been heard of since except in the rumblings of the IM community!
4) LSI was suppose to work at the level of “meaning” but there are just to many combinations for any algorithm to handle it!
5) Test it yourself, in LSI your search for a singular and plural word should be the same as they “mean” the same thing. Same thing for past, present and future tense of words. Even Synonyms.
6) Bottom line: LSI is not in place.
So what does this all mean?
1) Go out and get links from ALL the authority sites you can! 🙂
2) Don’t worry so much about the theme of a site that links to you.
Oh, Happy Day!
Nice post, and yes LSI is snake oil! Another one recently dismissed was the whole page rank sculpting using nofollow. Looks like good old fashioned flat site architecture, naming files and directories with keywords, title tags, and on page content are the still the pillars!
You HAVE to tell me if my trick worked on recovering PR from parked domains!! Email me. 🙂
nice post man! i will definitely be testing LSI. thank you for all this good info, you have alot of good post, and i am planning to read them all!
nice to read from your blog,but provide more information about LSI and i will definately test it
The whole idea of LSI really needs Web 2.0 to turn to Web 3.0 or even 4.0 before it has enough rich ontology content to even begin Semantic Indexing in any powerful way.
You are right, it is snake oil – now. Just because Google does not appear to have a patent there is no reason to believe that they do not. Google are good at search – they also have a nifty understanding of how to hide things. You could try looking at an alternative search engine to see what I mean.
I can able to understand LSI theory for some extent.What I understood is nice.Will u provide some more info?