jump to navigation

SES 2008 San Jose August 18, 2008

Posted by simarprit in : Bay Area, Blogging, Internet, Marketing, Search Engines , trackback

For all my posts on SES 2008: http://simarprit.wordpress.com/category/ses-2008/ 

Long Tail of Search & Semantic Search

It is my first time at Search Engine Strategies 2008 – San Jose, California (SES 2008, San Jose), two sessions down and currently attending the panel discussion here are my early thoughts on what’s going on:

Random Gyan:

Long Tail of Search

  1. Long tail can count to about 30% of the organic search
  2. Use blogs to harvest long tail traffic
  3. Use local long tail terms in local listings
  4. Use attributes in Google Maps
  5. Look for the “very Long Tail” and not just the long tail
  6. Weigh your advertising options
  7. Smaller Niche sites maybe a cheaper option, and may give great results at fraction of cost
  8. Links with navigation can help
  9. SEO’s have a role in online search market
  10. Mind your Traffic vs. Bounce rate
  11. Three phrases per page maybe the most optimum and effective targeting
  12. Live out of logs, they hold he key
  13. Shoot upto 20 “me too”: terms per key phrase
  14. Make sure your pages get indexed, search engines are busy, they may miss you.
  15. Divide your site into Key Phrase Zones
  16. SEO is no longer dead, it is making a comeback, and a big comeback
  17. PPC is becoming expensive
  18. Understand your Page Yield
  19. Account for your bottom 100 pages, analyzing them may help
  20. Attack your freeloaders
  21. Plan your SEO Dashboard, ensure it has everything you need, right under your mouse
  22. Measurement is critical, Measure visitors per page
  23. Know the lifetime value of your customers
  24. Search friendly URLs are important, but there is also life beyond them.
  25. Use “No Follow” effectively to increase your page rank
  26. Slicing and Dicing the content is critical
  27. Ideal length of the URL is a very subjective thought
  28. Anchors within the documents help
  29. Rate your keywords and assess your value
  30. Optimize Pagination
  31. Re-look at your internal linking structure
  32. Surrogate sites can help
  33. Getting high value back-links remains critical

Semantic Search:

  1. Let us talk to search engine and forget abstract codes and mono syllables
  2. Semantic Search engines attempt to understand underlying structured data
  3. Semantic search engines tend to eliminate ambiguous queries
  4. Search monkey project by Yahoo is an attempt to open up Yahoo Search platform for further development, based on semantic qualities and natural language expressions.
  5. Boorah claims that semantic search is happening. A quick 20 word analysis revealed little and exposed lots of limitations to the project.
  6. Google is implementing high degree of behind the scene semantics
  7. Hakia is a full blown version of Semantic Search (Beta), some of the attempts looked to be right, results are coincidentally close to Yahoo
  8. Powerset claims to have extreme semantic capabilities, which were not visible when I gave a set of questions to the website
  9. Ask.com claims to be amongst the best semantic search engines around with best results in Natural Language Query processing
  10. Generalization Vs. Specialization, Parallelization, Question Type, Categorization, Compressions, Content Characterization, ontology science applications are amongst some of the terms introduced.
  11. Inside the page semantic results could be another deliverable.
  12. Semantic Search has a lot to do with Aggregation of facts
  13. Improving relevance would remain a challenge
  14. Linguistic capabilities aided in ample measure with AI can deliver a great product over couple of years.

Comments»

1. K.N. Ajit Narayan - August 28, 2008

Hi Simar,

I wanted to post this comment earlier, but wanted to confirm a few things.

A month or two back, one of my friends who is pursuing his post graduate studies in Theoretical Mathematics explained to me an interesting concept of vagueness and completeness. After reading your post, it seems that Semantic Search is in some ways connected to the “Logic of Vagueness”.

Vagueness is something that human beings understand, but we haven’t been able to incorporate that ability into machines.

Without going too much into the mathematical details of the Logic of Vagueness, i’ll just speak about how it’s related to Semantic Search.

If you type “Gandhi’s assassin” into Google, the first result thrown up is a Rediff article; when ideally the Wikipedia page on Nathuram Godse could have come up at the top.

Currently, all search engines follow a statistical method of throwing up relevant URLs, wherein the URL that has used the keyword “Gandhi’s assassin”, the maximum or optimum number of times, occupies the first position in the search engine results.

According to the Logic of Vagueness, a human being will understand from the query, “Gandhi’s assassin”, that information about Nathuram Godse is being sought. A human being will then proceed to provide all information about Nathuram Godse.

If and when Semantic Search becomes a reality, a search engine would be endowed with the capacity to understand that when a person uses the keyword, “Gandhi’s assassin”, he or she is actually looking for information on Nathuram Godse. The search engines would then probably show the page on Wikipedia as the first result for the keyword.

I have taken up a lot of space, sorry about that…But a lot still remains to be said about this topic. Maybe some other time…