Skip to Main Content

Semantics for Search – Interview with Michael Marshall

Posted on 8.12.2010

Website Magazine contributor and co-founder and director of Search Engine Academy John Alexander interviews Michael Marshall on the topic of semantics and Natural Language Processing.


J. Alexander: Could you please briefly outline your background and qualifications for our readers?

M. Marshall: I have been in IT for over 20 years and in Internet Marketing since 2002.  I have degrees in Philosophy and Linguistics and background as a developer in Artificial Intelligence.  I have also been a consultant to the U.S. Patent Office, teaching courses in Search Engine Technology.  I have consulted for small, mid-size, and large companies including Fortune 500 companies and Internet Retailer 500 companies.

J. Alexander: For the sake of this interview, I'd like get you to share a little on the topics of semantics as it applies to search. Would you care to make 2 or 3 points that our readers may not be aware of concerning Natural Language Processing?

M. Marshall: 1) NLP is predominately a page-specific influence and, therefore, does not require webmasters to focus the content of their entire websites on 1 or 2 topics.  2) NLP looks at the context of a page, not just the target keyphrase in your copy and so cannot be tricked by techniques like keyword stuffing, repeating the main phrase over and over again.  3) NLP has a competitive component.  It is not just how well you might use thematically related terms, but also how well your content is written as compared to that of your competitors.

J. Alexander: Thanks Michael. I know you speak at many different events besides your own hands-on SEO classes and you have been very helpful to all our Search Engine Academy associates in developing many of the advanced lessons.

Thinking back over all the classes that you have taught, what might you say the single biggest surprise is to those just trying to grasp NLP and how influential it is?  

M. Marshall: The biggest surprise tends to be how learning to writing content in a way that takes into account the influence of NLP also helps address an important aspect of the challenge posed by the personalization of search results; so that high visibility for various segments of your target audience can still be maintained even after personalization shuffles the search results.

J. Alexander: We all hear terms which are used or sometimes confused or sound confusing because sometimes no one has really thrown much light on these areas. Talk to us a little bit about how some of the search engines appear "to think." We all know they do not have human intelligence. So give us a factor or 2 that gives the search engine the appearance that it can think.

M. Marshall: One way a search engine appears to think is in its implementation of contextual advertising.  Based on the content of the page you are viewing, it serves up ads that are thematically related to the page.  The relevance matching, when it works well, often looks like human understanding of the content, whether it is based on complex keyword matching or more sophisticated systems that use a language-independent proximity pattern matching algorithm to increase matching accuracy.  

With any of the functionality a search engine has that give the appearance of thinking, it is predominately the technologies it uses for various forms of sophisticated pattern matching and multivariate analysis which make such functionality possible.

J. Alexander: Thank you. Is there anything else you care to comment on with respect to artificial intelligence and search engines.

M. Marshall: When discussing this topic in the SEO community, usually LSI (Latent Semantic Indexing) is the focus.  LSI is only one of many types of NLP, but LSI often gets misunderstood in the SEO community as the only means by which a search engine might use AI to better ascertain the meaning of a document.  This, in turn, often leads to viewing arguments against the belief that search engines use certain forms of LSI as some kind of demonstration that no semantic AI at all is at work or has influence on search results, when there are a myriad of other NLP techniques available.  

J. Alexander: Thank you very much Michael Marshall for your time today answering these questions and helping readers gain a better understanding of semantics, artificial intelligence and Natural Language Processing. Good luck with your hands-on classes at Search Engine Academy North Carolina.

 

 

About John Alexander: John Alexander is co-founder and director of the Search Engine Academy, leading provider of hands-on search engine optimization training at locations across the United States, Canada, Asia and Australia. To find an SEO workshop nearest you, visit Search Engine Academy

About Michael Marshall: Michael Marshall is CEO of Internet Marketing Analysts, LLC, a provider of SEO consulting services and highly acclaimed SEO competitive intelligence tools.  He is also the owner of the Search Engine Academy of North Carolina.

 

Leave Your Comment

Login to Comment

Become a Member

Not already a part of our community?
Sign up to participate in the discussion. It's free and quick.

Sign Up

 

Leave a comment
    Load more comments
    New code
  •    

    The Ultimate Guide to Personalization

    Kibo