Page 1 of 3 123 Last
  • Jump to page:
    #1
  1. No Profile Picture
    Contributing User
    SEO Chat Good Citizen (1000 - 1499 posts)

    Join Date
    Apr 2004
    Posts
    1,179
    Rep Power
    17

    Google's 2005 Superbowl Update - the Whole Picture


    What happened on Super bowl update

    from syntactic match to semantic match - Latent Semantic Index (LSI) or Latent Semantic Analysis (LSA) is a well-known technology
    in IR (Information Retrieval) field. However, it's not important to know (we won't know) what specific technology Google uses for
    semantic analysis. What important is that we see certain degree of semantic match. Search terms may not appear in the title or even
    body of a document that is ranked high on SERP, but we can find semantically related words in the titles or those document.
    The implication for SEO is simple and straightforward.

    from PageRank to SiteRank - What we witnessed, after the update, is that many sites entirely dropped off SERP. There're no other
    explanation other than criteria at site level are applied in this update. I'll call this SiteRank. If PageRank measures the
    importance of an individual page, SiteRank measures the quality of a site. Major factors that can and should be used include,
    not limited to, 1) strength of content (size of the related content pages), 2) quality of links (links from diversified sites with
    variant anchor text to many different pages), 3) freshness of the content (regular update of content), 4) uniqueness of
    content (less percentage of duplicate content), 4) age of the site, 5) outgoing links (less percentage of deadlinks and more relevant links),
    and 6) a reasonable Pagerank. If you have a lower SiteRank, you can't even find your site using your company name, instead you found websites that
    link to your site.


    Philosophy behind Super bowl Update

    It is the philosophy, not simply the technology, that will distinguish winners from losers in the search market. I'll briefly
    examine the philosophy behind the Superbowl update.

    searcher engine can improve quality and relevance by tweaking algorithms - This is only partially true. In many case, users
    can't find what they want is not because search engine aren't smart enough, but simply because these pages don't exist at all.
    Majority of Web documents is written for sequentially access (reading a book or a newspaper), not written for search (or random access at all).
    Good search engines should encourage publishers to create more pages for searchers.

    SEO does no good for search engines - This is not true. The problem is that search engines can't understand the content and
    evaluate the quality of a document. Instead, they use link analysis to proximately do the job, sometimes, very poorly.

    Semantic Analysis will improve search quality - Well, may be. Strictly speaking, LSI isn't a science or theory, it's a technology.
    The success of semantic analysis is based on the assumption that semantic relationship exist with a document set. This may be true for
    certain types of content.

    quality site will produce quality document - Not true. Many useful and insightful pages are created by individuals
    who are experts in their domains, and don't have or don't want to access to so called authority sites.


    What's Next


    It's premature to speculate when and what will happen for next major update. It's gonna be another 10 -14 months away from
    now. There may be on-going debate within Googleplex for the directions of next major update. But for upcoming monthly minor update,
    we can expect Google to go in the same direction of Super Bowl update to further refine and stabilize the major change with the ideas
    of placing some focus on semantic match and site quality. I've not completed a thorough and scientific test of new search results, the primary
    impression was that the update has improved the relevance and quality of more general search terms (1 or two word search
    queries), but it's harder to find specific information using specific search terms (3 or more word search terms) which was
    Google's major competitive advantage over Yahoo and MSN. I'd expect Google to place less weight on semantic match and
    site quality at least for more specific search terms in coming monthly updates.

    Bruce

    This is part of an article I wrote today. Feedback welcome. The full article is at: http://www.4th-media.com/google/superbowl_update.php

    Comments on this post

    • gchaney : It is really inappropriate to post theory as apparent fact. It can mis-lead those who know don't no better.
    • gbull
    Last edited by dejaone; Feb 22nd, 2005 at 04:23 PM.
  2. #2
  3. No Profile Picture
    Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    May 2003
    Posts
    282
    Rep Power
    17
    Originally Posted by dejaone
    Major factors that can and should be used include, not limited to, 1) strength of content (size of the related content pages), 2) quality of links (links from diversified sites with variant anchor text to many different pages), 3) freshness of the content (regular update of content), 4) uniqueness of content (less percentage of duplicate content), 4) age of the site, 5) outgoing links (less percentage of deadlinks and more relevant links), and 6) a reasonable Pagerank.

    Most of contradicts what happen to my blog. I had a PR7 site before the update (its now PR6), the blog is unique and has fresh content, my site has been indexed by Google since nearly Google's beginning, outgoing links are relevant and not dead. Yet, my blog got trashed and now sits behind mutliple blogs with no PR, that aren't as relevant, and haven't been updated in ages.

    In addition, none of this explains the drop in relevancy, why spammy sites have improved/held their rank, and sites that are merely index templates continue to do well in the rankings.

    With all do respect, I have yet to see across-board-the-evidence that supports your theory.
  4. #3
  5. Free the SB
    SEO Chat Skiller (1500 - 1999 posts)

    Join Date
    May 2004
    Location
    DC region
    Posts
    1,832
    Rep Power
    21
    Bruce:

    I've been searching for coherent descriptions of the impact of the Super Bowl update. It impacted my serps and those of competitors plus sites now in development. It essentially improved the positioning of direct competitors who were nowhere to be found on google's earlier rankings. Frankly for some of the terms I follow I think Googles serps are better, regardless of the fact that my site went up and down in a bit for different keyword combinations.

    Where is your information coming from. I've been reading through many threads with various opinions but not getting a coherent sense as of yet as to the impact. (Maybe it is too early). Where do you see differences between simpler and more complex terms. I'm eager to hear more.

    Dave
  6. #4
  7. SEO Chat Skiller (1500 - 1999 posts)

    Join Date
    Jul 2004
    Location
    Seattle, WA
    Posts
    1,888
    Rep Power
    47
    Bruce,

    Your paper suggests a lack of understanding of semantics, LSA/I and the impact on search technology. I'd urge you to be more knowledgable on the subject before including these types of assertions in such an authoritative way.

    I barely know enough about semantic connectivity in search to talk about it and it has taken months of reading and digesting with help from some skilled professionals.

    I'm also at a loss to come away from your paper with conclusions about the update. You suggest that PageRank (global link popularity) is less important than something you've aribitrarily dubbed "siterank" which includes many supposed "quality" measures. I'll often come up with my own names for theories, but the problem is that you've grouped many known SEO factors into one lump term and left it open so that any qualititative measurement fits in.

    Hope to see more writing from you in the future and don't worry - we all start somewhere, perhaps there are others who can take away more from the article than I can.

    p.s. If you could edit your post's spacing, it would make it much more legible. Thanks!
    CEO & Co-founder of SEOmoz
  8. #5
  9. No Profile Picture
    Contributing User
    SEO Chat Good Citizen (1000 - 1499 posts)

    Join Date
    Apr 2004
    Posts
    1,179
    Rep Power
    17
    Sweat, Thanks for the feedback.

    One important thing I've not explained in the post and the article is the nature of SEO theories (if any).
    Search engines use stochastic models to rank search results. Meaning that good sites could be buried by good (at least with good intention) algorithms.
    Whether one explanation offer valid insights, we have to look at the percentage of sites that can fit into the patterns.

    I'm in the process of doing some thorough test to see whether my ideas are statistically valid.
  10. #6
  11. No Profile Picture
    Contributing User
    SEO Chat Good Citizen (1000 - 1499 posts)

    Join Date
    Apr 2004
    Posts
    1,179
    Rep Power
    17
    To answer your question, Randfish. I've read all original LSI papers back in early 90s. They're mathematically simple for someone who was trained and experienced in mathematics. I even wrote a small program that implemented LSI. I do appreciate your other feedback.
  12. #7
  13. No Profile Picture
    Contributing User
    SEO Chat Good Citizen (1000 - 1499 posts)

    Join Date
    Apr 2004
    Posts
    1,179
    Rep Power
    17
    Dave,

    The ideas are from primary obervation + speculation. I'm doing some thorough and scientific test. The test may prove or dis-prove my ideas.

    Bruce
  14. #8
  15. No Profile Picture
    Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    Jan 2005
    Posts
    193
    Rep Power
    15
    Whether it's LSI/LSA or something else, I feel like this isn't going to be good for Google in the long run, if this is indeed the intended outcome. When people search for a Blue Widget, they don't care about the process Google goes through to show them results, they just want to see relevant pages.

    If the pages they get leave them feeling somewhat empty, they're going to go look elsewhere. And when they find just what they wanted at #1 in Yahoo or MSN, they're going to wonder why that page wasn't in Google.

    That's one demerit point for Google. If this keeps happening, they will tell their friends. And what once made Google popular (relevant searches) will now make another SE popular.

    I understand the basic concept here, but I really don't think what Google has right now is what they wanted - or at least I should say, I HOPE they don't have what they wanted!

    Drew
  16. #9
  17. Free the SB
    SEO Chat Skiller (1500 - 1999 posts)

    Join Date
    May 2004
    Location
    DC region
    Posts
    1,832
    Rep Power
    21
    Good luck Bruce:

    At this point I think its hard to define the impact of this dance. I've been through many forums and have yet to hear a defining set of comments or description of the dance. I've been through some of the threads on content issues like LSA (totally new to me) and while I find these discussions illuminating and compelling I have yet to see evidence of exactly how this has played out.

    I've been looking through a variety of search phrases for some similar types of topics where pre-Feb 2, the highest ranking serps were comprised of providers of certain services and directories or well optimized commercial descriptions of these services and the service providers themselves.

    Post Feb 2 I see a shake-up but it is not consistent. If anything there are different types of reshuffling of serps somewhat contingent on the overall level of competitiveness.

    I'll be interested to see what you find.

    Dave
  18. #10
  19. No Profile Picture
    Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    May 2003
    Posts
    282
    Rep Power
    17
    Bruce,

    While I may disagree with your article, I appreciate the fact that you are coming back to answer criticisms of your theories and appear willing to listen to other viewpoints.

    If you do more research, please comeback here and post the results (or at least a link to the results).
  20. #11
  21. No Profile Picture
    Contributing User
    SEO Chat Explorer (0 - 99 posts)

    Join Date
    Nov 2004
    Posts
    57
    Rep Power
    15

    Question Should we start?


    I think that, regardless of this discussion, LSI is having SOME effect on the latest Google changes.

    Should we start applying some LSI optimization to our websites?

    For instance, I wonder if to target "Cheap Sunglasses" we should change a title like:

    "Cheap sunglasses - XXX sunglass company"


    to

    "Cheap sunglasses and lenses - XXX sunglass company"

    not to target all these keywords but in order to use keywords in the same semantic group.-
  22. #12
  23. No Profile Picture
    Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    Aug 2004
    Posts
    179
    Rep Power
    15
    Originally Posted by Renton
    I think that, regardless of this discussion, LSI is having SOME effect on the latest Google changes.

    Should we start applying some LSI optimization to our websites?

    For instance, I wonder if to target "Cheap Sunglasses" we should change a title like:

    "Cheap sunglasses - XXX sunglass company"


    to

    "Cheap sunglasses and lenses - XXX sunglass company"

    not to target all these keywords but in order to use keywords in the same semantic group.-

    This is a pointless topic. Generally, people could care less. They get listings based on what they type in to the search box. Do you think average Joe Blow gives a sh!t about how these results came to be.

    Do you think the average Joe Blow thinks that these are not the best results? Don't think so.

    If your not happy about your SEO strategy, tinker with it. Can't hurt you unless your top 20 and then i guess you wouldn't have even wrote a theory to begin with.
  24. #13
  25. SEO Chat Skiller (1500 - 1999 posts)

    Join Date
    Jul 2004
    Location
    Seattle, WA
    Posts
    1,888
    Rep Power
    47
    Let's be careful about the use of the acronyms LSI/LSA. This is a specific method of analyzing semantic relationships, not the concept overall. Glad to see your thinking moving in this direction, though - check out the advanced on/off-page optimization thread.
  26. #14
  27. No Profile Picture
    Contributing User
    SEO Chat Explorer (0 - 99 posts)

    Join Date
    Nov 2004
    Posts
    57
    Rep Power
    15
    Originally Posted by jrath1
    This is a pointless topic.
    I don't think this topic is pointless. Many of us (most of us, IMO) have been severely damaged in some way by the latest Google update.

    And i'm not seeing many theories regarding what happened and what we can do to get our rankings back.

    In consecuence, I wonder if having keywords in the same semantic group of the ones I target on my website will help getting my ranks back.

    This adds up to some other theories, like the one of a new overoptimization filter, "too many, too soon" links filter, etc.

    This are all theories, meaning they arenīt certain.

    But what I think is pretty certain is the fact that this G update is here to stay, so we better start acting about it.
  28. #15
  29. Contributing User
    SEO Chat Discoverer (100 - 499 posts)

    Join Date
    Oct 2004
    Posts
    121
    Rep Power
    15
    Originally Posted by Renton
    But what I think is pretty certain is the fact that this G update is here to stay, so we better start acting about it.
    Are you really certain? I'm not sure it is. Isn't there still a lot of activity in the DCs? I think anything is possible at this point.
Page 1 of 3 123 Last
  • Jump to page:

Similar Threads

  1. Google Dir Update! (Feb 8, 2005)
    By Mok in forum Google Optimization
    Replies: 0
    Last Post: Feb 8th, 2005, 12:36 PM
  2. superbowl update regional search?
    By jwbond in forum Google Optimization
    Replies: 8
    Last Post: Feb 7th, 2005, 04:52 PM
  3. Google's directory updated!
    By Mok in forum Google Optimization
    Replies: 33
    Last Post: Jan 17th, 2005, 05:44 PM
  4. googles next pr update
    By wraptduck in forum Google Optimization
    Replies: 10
    Last Post: May 27th, 2004, 09:12 AM
  5. Cached pages, deep crawl and googles update
    By jungleimp in forum Google Optimization
    Replies: 1
    Last Post: Mar 27th, 2003, 06:40 AM

IMN logo majestic logo threadwatch logo seochat tools logo