LSI
- Latent Semantic Indexing to become
an even more important Future influence
by
John Alexander
It was back in a Vancouver Workshop
a few years ago, that we had invited Michael Marshall, an expert on Latent Semantic
Indexing to come teach our students on
the subject. Few people were discussing this topic back
then.
As Michael began to teach the class, Robin Nobles
looked across the aisle at me and I glanced back at her as if
to say....where in the world does this fellow gather his
insights from? Robin already anticipated some exciting things
to come because she had known Michael as one of her early
online SEO students. I had only just met Michael but was quite
impressed with not only the range of detail he presented, but
also the way he presented it, as he explained LSI
and literally drew mathematical diagrams on
the board. He worked away at explaining term vector databases and
how artificial intelligence was working within some of the
search engines like Google. Of course one of the advantages
that Michael has as a communicator, is that he is also an expert in
Linguistics and also in Mathematics and. He was able explain and
unravel all of the mysteries of LSI on a level that all of our
students could understand. But even better yet, he helped teach
them simple ways to take advantage of these influences without
even needing to be a math expert.
Back then, we had already anticipated
some new rifts and changes in importance of this type of
influence, since the results of those early tests still defy
some peoples reasoning, but for those who were in the know,
the knowledge brought us and our Workshop students some significantly
wonderful successes.
Let's go back a bit to the beginning.
What is the study of "Semantics?"
In general, or in the simplest
of terms, "semantics" is the study of
"language meaning."
Here is a quote from Michael Marshall and how he
describes "semantics" and specifically
how it applies
to LSI on the Web:
"The field of semantics includes
the study of how meaning is constructed, interpreted,
clarified, obscured, illustrated, simplified, negotiated,
contradicted and paraphrased. In other words, one might say
how words relate to other
words in
order to put the meaning of a message into the correct
context.
With respect to LSI, semantics deals with how the various
aspects of meaning outlined above, can be found "hidden
in documents" by means of the presence, absence,
associations and relationships between words found in those
documents."
Michael Marshall
Internet
Marketing Analysts
One of the most important
things to understand is that Latent Symantec Indexing is NOT
a replacement for search engine optimization. It is not the
latest thing to go chasing after and drop everything else
you have learned.
Also remember that it really is nothing new and has been
around for a few years. We've been teaching strategies that
employ these influences to our Workshop attendees for over 3
years now. But Latent Semantic Indexing along with it's
acronym (LSI) certainly has become one of the current industry "buzz
words," since many people are just recently learning
about its existence, its strengths and the impact it can
have on building stability in your rankings. We've
also seen instances, where some people are using it as a new
"buzz word" but have not really worked with it first hand, to
actually test it out, understand it or truly make the best use of it.
We have tested it out and it is still very valid as just one of the advanced
SEO strategies that we teach at our workshops today. We have
also had many of our research team conduct tests. Our
students themselves have also reported great success
by employing these strategies that apply a knowledge of LSI
into the balance of optimization factors we have taught
them.
Traditional Search Engine Marketing
One of the first things you learn about in structured optimization
training, is that you must be very focused on a specific
keyword phrase that you have already properly researched.
LSI though is not so much about your researched keyword
phrase, but its much more about the body of surrounding text
that gives meaning to the document as whole.
It is true that a
human being can scan a page and by simply reading it (with human
intelligence) we can read the page and very easily determine what a page is about.
However, without applying traditional optimization
principles and even though we can understand what a page is
about, the page still may not be ranked unless we apply a
range of optimization factors. How hard you have to work to
achieve those rankings, will depend on how competitive your
phrase is.
Results of Early Testing of LSI
Suppose you saw a page with literally no optimization
factors being employed. No on-page influences. Literally no off-page
factors at all. No "linking strategies" or
anything that you are ordinarily accustomed to seeing in
traditional optimization practices. In other words no
existing "white hat" optimization and no tricky grey area
or "black hat" optimization stuff, no
cheats, no cloaking or no Spam. And yet suppose such a page
was still deemed more relevant for a highly competitive term
than literally 148 million other pages, all that were
properly optimized.
The question would remain, how could such a page be ranking
over multi-millions of competitors for and extremely
competitive term? As a human being reads the page of
content, of course the content makes logical sense and has
value. But how on earth would a search engine possibly
"understand" the context of the overall page and
give it a top ranking spot over millions of competitors when
it is using literally zero optimization practices in
place?
Its scary to the point where you would want to wonder how
such a page could be honestly out-ranking other pages that
are highly optimized, high PR and even have millions of
inbound links.
Could it be that a search engine really can
"think for itself" and absorb the value of the overall page?
So the question is...
How can a search engine "appear to be able to
think?"
Although the LSI algorithm doesn't really understand
anything about what the specific words on a page mean (after
all a search engine cannot really think), but the patterns
it notices through Artificial Intelligence or AI can make it
seem extremely intelligent.
At our advanced 3-Day Workshops, we demonstrate to our students just how powerful the LSI influence is by showing
them some of our early tests which are still pretty amazing to
examine. At the same time, our whole focus is to teach each
of our students the full scope of skills and remind them of the
importance of building "genuine relevancy" for
"useful content" that has true value to a visitor and merits top placement.
Why LSI will continue to become more important in the
future.
One of the main reasons that Latent Semantic
Indexing will continue to be an important future influence
is because LSI is actually
rewarding relevancy to documents that are well-written and
truly meaningful to a human reader. As I said earlier, a search
engine cannot really "think" but it can appear to
think. More and more, search engines are beginning to
recognize the context of a written message.
Latent semantic indexing actually adds an extra step to the
page indexing process. In addition to recording which
keywords that a specific page contains (as with traditional SEO),
the LSI influence examines the document collection "as a
whole" to see which other documents contain some
"Semantic similarity."
Human understanding may differ from what a search engine
"understands."
Remember that we can not completely base the relationship between
different words and synonyms only on what we as "human
beings" know. LSI is about what a search engine
"understands" through it's entire collection of
data. But thanks to AI and LSI at work in the background, the method
does correlate surprisingly
well with how a human visitor, who is looking at the content, might
decide to classify a document collection. In short, it
appears pretty intelligent.
Remember that there is a lot more documents being indexed
and included in a search engine's database thank just
typical Web pages. Think about the huge impact of news
stories and other information being published for example on
Blogs. Do you think that if a certain public figure or a
celebrity was getting a lot of bad press in the news, that
that gathered data might not also influence which very
specific words might now relate to that persons name, based
on a search engines "understanding?"
Great future rewards for professional writers but bad
news for Spammers.
This is another reason why many well written articles that are published on
the Web do extremely well in the rankings. It's because they
are "well written" by writers who are not even worried
in the
least about what their keyword density, keyword prominence
or the rest of the hundred and some odd technical elements
that optimizers tend to fret about.
The writer is focused on
delivering something meaningful to their readers. This is
not to say, that there is no place for the technical aspects
of doing good optimization work, because it is still very important.
But
optimizers need to understand the value and the rewards that
go along with creating content they can be proud of. Its no
longer just only about optimization in the sense of top
rankings, but its about providing fresh, original content
that has true value to a reader. The day may come when the
first search engine will actually "win the war"
on Spam. Whichever search engine that manages to achieve
that
first, will manage to also become the leading search engine of their day.
The public wants useful,
quality information. They want it nicely delivered with
speed and minimal fuss. Everyone wants more accuracy in
search. Everyone is tired of clicking on unhelpful
or useless pages that have no value.
Thanks to LSI and theme based search engines that using a
technology that employs Semantics, LSI, and AI, the Webmasters who continue to focus on creating high quality,
well written and engaging content, which "reads well" to a human being, will be better rewarded with higher rankings,
better visibility and ultimately much more business and
sales too.
How up to date are you on all of the SEO influences that are
currently "working in the now?"
Check out SEO Mastery Workshop future
dates and locations here.
Check out the Complete
5-Day Workshop agenda
Grab
a Free SEO Tip and have it sent to you by e-mail every
day - hundreds of free tips!
Subscribe or unsubscribe as often as you wish.
About John Alexander
John Alexander is Co-director of Training at Search
Engine Workshops offering live, SEO Workshops with partner
Robin Nobles as well as online search engine marketing courses
through Online Web
Training. John is author of an
e-book called Wordtracker
Magic and has
taught SEO skills to people from 87 different countries.
John's articles can be read in publications like REALTOR
Magazine, Search
Engine Guide,WEBpro
News, SitePro News
and many others.
Click here to go back to the
index of search engine marketing articles
This work is licensed
under a Creative Commons License.
We'll keep you advised on all of our future
training dates the moment things are firmed up.
Simply enter your name and e-mail in the box below.
See
what other dates and locations are still available
|
John Alexander
|
