July 11, 2007

Search: Finding or Understanding

Many of the researches have shown the increasing trend in the usage of web search. Since more and more people are using search to find various web addresses, the expenditure on SEO and SEM has also increased many folds. Even the Non IT Companies are trying hard to make them easily searchable on the popular search engines.

But have we ever thought that does the search engine understand what we want? The answer will be “No” in almost all the cases. Most of the search engines will just simply do a match of your query in their database and will give you all the documents matching your search criteria. But what differentiates a good search engine from an ordinary one, is the ordering of those results. Thus the main expertise (apart from having a big database) lies in how to show the most relevant results first to the user, out of thousands of matching results.

So all the googles in this world started focusing more and more on how can they improve the ordering of the results and came up with various complex algorithms. But still all they can do is “finding” and not “understanding”. For them, what you are looking for are merely “keywords” and nothing more than that. I think I will be able to explain it better by this example.

Let’s say you are in love with your fishes and you are looking for some healthy food for them. Someone suggested “just Google it”. So you wrote healthy food for fish on the magical text box and pressed go, expecting that it will tell the best possible food for your fishes. But you don’t know that “for” can be a stop word in Google and using its world famous page rank algorithm it will search for “healthy”, “food” and “fish”. So you will get nearly 3,230,000 results and but you will be disappointed to see that most of them are of “healthy fish food” and not “healthy food for fish”. And after going through all the URLs, I won’t be surprised if you end up eating all the fishes in you aquarium .

But it’s good to know that some people are working on various concepts like web2.0, web3.0, semantic web and so on, to improve the relevance of their search results. Some of them have started using the user himself to build intelligent algorithms. But still I don’t know when will they be able to “understand” what the user really expects of them?

No comments: