Hi. This is an old, unmaintained blog. You may find these sites more to your liking:

Carson Brackney: This is my primary site.

Ad Astra Traffic: Content production/article writing service.

Ad Astra Traffic Team: For those who'd like to get writing gigs with Ad Astra.

Monday, October 30, 2006

Getting it backwards on latent semantic indexing and online content...

Google uses latent semantic indexing (LSI) to help determine search results.

Thus, some SEO "experts" and content providers are advising webmasters to use LSI-compatible or LSI-friendly text on their sites to improve SERPs.

They have it backwards.

Let's start at the beginning. Latent semantic indexing, according to Dr. Ralph Wilson, "is an approach to understanding keywords in the context of the words on the entire webpage." In a sense, it is a mathematical approach to interpreting content based on expected themes.

It offers real value to the search engines. If you have a page that keeps mentioning "cards," for instance, LSI will help Google determine if you are writing about "playing cards," "greeting cards," "business cards," or the "St. Louis Cards" by evaluating the surrounding content. LSI methods should help create a more accurate means of information retrieval. It's also a big part of the reason why you can get a great result for an "automobile" page even when you used the word "car."

To understand why the people hyping "LSI content" have it backwards, you need to consider the impetus behind Google's decision to use LSI and, more importantly, their primary function for existing in the first place. As much as some folks in the SEO and content industries would like you to believe otherwise, Google's #1 mission is not making it more difficult for you to rank for your keywords. They bought CIRCA and its LSI methodology in order to produce more accurate and meaningful search engine results. Their primary function is to serve up good results to users.

So, what constitutes a good result? Putting the information people really want on their monitors as quickly and easily as possible. If Google can broker successful searcher/information deals, it will continue to flourish. If the SERPs becoming meaningless or get too crummy, Google is in trouble.

Thus, all of the algorithmic tweaks Google undertakes are designed to keep the SERPs clean and meaningful. They change the secret recipe in order to improve the overall quality of the searcher experience and to put people and information together as efficiently as possible.

It's easy to see how latent semantic indexing can help, isn't it? Google can give you the recipe for a great homemade catsup even if you were looking for "ketchup." Google can better "understand" context, which can prevent it from sending out incredibly off-base search engine results.

Some people have maintained that Google is also using LSI as a filtering technique. They might claim that G checks for your keywords and if it DOESN'T find corroborative terminology via LSI, you will be punished.

I am not going to claim to be the world's foremost "nuts and bolts" SEO expert. I have no idea if that is true or not. Based on what I have read from sources I tend to trust the most and from what I've been able to glean from some pretty compelling, more academic assessments of LSI and information retrieval, I have my doubts.

Thus, the "you need LSI-ready content" arguments are, in all likelihood, bankrupt. However, even IF Google made significant use of LSI as a means of downgrading pages that didn't seem to make sense "contextually," that isn't a reason to buy content from someone who claims they have somehow figured out the latent semantic indexing formula and are capable of providing writing that will "pass through" some ostensible "LSI filter."

That's because LSI, and everything else the search engines do in terms of algorithms, is designed to put good information and users together. That's still the bottom line, as we discussed. We know a few of the reasons why Google uses LSI and we may or may not believe the idea that it uses LSI as a mechanism for punishing contents that lack expected surrounding context. Either way, the goal remains the same: Get good information on top of the SERPs.

People continually fail to recognize that Google's algorithm isn't a game for its own sake. It is a device designed to automate a uniquely human process--information assessment. The best way to fall into G's good graces as those efforts are improved and fine-tuned? Good content.

Guess what? If you hire a good content writer to prepare an article on a particular subject, he or she is going to produce and informative, factually accurate and readable piece on that topic. He or she will naturally use appropriate synonyms and contextually relevant terms and phrases. The material will flow through any real or imagine LSI filter easily because it will be good, not because someone tried to reverse engineer the LSI system.

If hire Content Done Better to write articles on a certain kind of automobile, you can bet your last dollar those articles will have the "right" associated words within them. That's not because I am a student of LSI or an algorithm code-buster. It's because I am a writer capable of producing quality web content. The right things are going to happen organically and the artificial technology of Google is just as likely to recognize that as it is to recognize some sort of "LSI-compatible" material.

Those who watch Google and then try to respond to its alterations are working backwards. Chasing Google is little more than stepping upon some kind of SEO treadmill. You look for angles, take them, Google adjusts for them and you start looking again. It never ends.

However, if you start where Google starts--with the goal of quality content and information, everything else falls into place. That "magic" keyword density (which, by the way, is bunk)? That use of appropriate phraseology in surrounding text to comport with LSI expectations? You don't need to worry about them. You need to worry about supplying good content that will actually match user needs. You can meet G's expectations and your own without worrying about KD and LSI.

Additionally, you gain the advantage of potential organic backlink creation (another staple of quality SERPs) with good content. Stuff people want to read. In fact, that's why Google values those links in the first place.

Chasing after LSI adjustments is just another example of approaching the content issue with respect to search engines backwards. Instead of worrying about how Google is trying to get better at finding quality content, consider simply dedicating yourself to actually securing good material.

This shouldn't be read as an indict of all SEO activity. There are things you can and should do to improve your search engine results. One of the most valuable ways--and one that allows you to avoid being sucked into crazy claims about minute algorithmic tweaks--is to invest in quality content.

Additional Reading from Randfish at SEOMoz.org...