Thummit does the Oscars

Search Results - A matter of Perception.

Interesting to note that there is really not much distinguishing the major search engines in terms of the actual search results returned. This article "
Google And MSN Tie, But Google Still Wins" gives details of a survey which indicates that although the users are generally satisfied with the results of the respective search engines, perception play a very large role in the search engine that they use. Google with it's towering brand equity in the search market has gained the most from this perception while MSN, though technically equivalent, I believe has been impacted negatively due to this. But it looks as if Ask Jeeves and MSN have made tremendous improvements and increase in their user base over the past year. This is also partly due to the growing market for search.

Information Overload

I have been hearing and reading a lot about this term "Information overload" recently. Just today, there was a post on Greg Linden's blog about it.There is no doubt that is a problem that is increasing in magnitude every single day, even as I write this post adding to the problem.

Most of the suggestions to solve this problem have been mainly related to better understanding the intent of the user, using the user's history and clickstream to further present the most relevant results. Google has recently started this and Findory has been doing this for quite some time. Although these work well in certain cases, I personally do not think this is the holy grail.

We here at i411 believe that there is only so much you can do to infer a user's intent. We prefer to give the user options and engage him in a dialog and give the user options so they can guide themselves to the right set of results. This approach works very well , especially when there is some structure in the data which you can take advantage of to guide the user. This is a trickier problem to solve for the web but not many people have attempted it. Vivisimo with their Clusty have attempted a solution with not much success and Google and MSN have also been researching with clustering technology but have not been able to bring it to the mainstream yet.

Some other potential combinations of solutions to this problem would also include

• New visualization techniques that would enable us to get a good sense of all the information presented to us quickly and intuitively.
• Authoritative source of information - The web has made each one of us a journalist and there is no pecking order among the sources of information. Maybe we need authoritative sources of information which take precedence.

An interesting quote to end the post -

"We are drowning in information but are starving for knowledge"

41% of Google Search Results Page are Ads

Someone did a pretty interesting analysis on Google's results page. We did not realize but this say's that Google is showing a lot more ads on their results page. They are now not much different from most of the other sites except they have managed to keep the interface clean.

Acknowledgments

I would like to acknowledge two bloggers that I have been following for quite some time and who have given me inspiration to start my own blog. I would like to acknowledge Greg Linden of Findory and John Battelle , the author of "The Search" .

Search Market Pie Expanding

According to this article, the search market pie is expanding rapidly and Google seems to be at the forefront grabbing those eyeballs. For the month of October, Google had 2.5 Billion of the total of 5.1 Billion search queries performed with Yahoo, MSN, AOL and Ask Jeeves making up the majority of the rest.
Although there is a big grab for the eyeballs for search, where are people headed with this ? As the competition heats up, the cost of these eyeballs is going to increase and the revenue for ads per eyeball is also going to decrease. Search advertising based organizations such as Google, where 99% of their revenue comes from their advertising need to seriously be thinking about getting more diversified sources of revenues. Not for lack of trying, Google has not succeeded in bringing a single block buster revenue making product to the market past its AdSense and AdWords products. With MSN, Yahoo, AskJeeves and a host of other startups on their tail, they are going to need to run faster than ever before!

Another Local Search Site !!

Here is the launch of another local search site -Tyloon. Their claim to fame is the ability to search local listings in Spanish and Chinese. Spanish I can understand but Chinese!! Why would someone in America want to search for businesses in Chinese ? I checked out the implementation of their search features and it is quite mediocre. They do not have additional content like some of the other new players (Local.com, truelocal.com ) and no mapping functionality as well like some of the other major players like Google, MSN and Yahoo. I tried putting in "Winchester VA" in the where box and it gave me zero results and if I just put in "Winchester", it have me results from at least 6 different states - not sure how useful that is for a local search functionality.

Couple of interesting articles on local search - 411 on local search and the role of IYPs in local search.

Enterprise Search - Coming of age ?

Here is an interesting interview of Torbjorn Kanestrom of FAST on Red Herring regarding the enterprise search search space. I personally like the analogy to the database market that he makes in the article. The example with respect to the the yellowpage companies using databases as surrogates for search technology is so true. As one of the major players in the yellowpages search market,we at i411 ( I work for them !) have encountered similarbehaviorr which has been changing over the past couple of years.
There isdefinitelyy going to be some consolidation in the search market over the next few years. One of the main things that will precipitate this consolidation will be the need for standards. Currently, every search engine has its own proprietary formats with no industry standards and no interoperability between them. The web services API effort from the major search engines is a step in that direction for web search and a similar effort will be required in the enterprise search space as well.
Interesting comment about R&D in the enterprise search regarding understanding content as well as query disambiguation. I think there will also be an acute need for new techniques in visualization as the amount of data in the enterprise grows and users are faced with information overload. One of the other major areas of research will need to be in thestructuree and format of the storage of data in the enterprise. Currently, enterprises use various methods ranging from CMS's, databases, etc to just pure file systems which makes the problem of enterprise search all the more difficult. Again standards regarding the storage of enterprise information will go a long way in helping to understand the content and its context.
As this is my first post into the exciting world of blogging, I would request the readers to provide feedback to help me improve my blogging skills.