Every data source, in one way or the other, has means to make searching for information easy. Libraries across the world use various indexes based on author’s name, book title, category etc. to make it easy for people to locate what they are looking for; making the search experience simple and relevant as much as possible. Internet – the largest and most widely used source of information has search engines to aid searcher in finding web pages matching their search query.
Search engines use complex algorithms to rank websites and present to the user web pages that are most relevant to his/her search. Relevance of search results is one thing that search engine constantly work to improve. This they do by adding new parameters and tweaking existing parameters to effectively judge a webpage and its relevance to search terms. This judgment is purely devoid of human intervention (at least for majority of the cases) because search engines spiders can emulate human behavior or so we are lead to believe. Spiders are touted to have the ability to perceive contextual references (Latent Semantic Indexing) and therefore to a great extent reproduce human intelligence. So does it effectively mean that human behavior has no significance in search engine rankings or should human behavior play some role (if not a major one) in deciding search results?
I am not debating whether or not search engines already use behavioral data to modify search results (as alleged by many and I do not intend to allay privacy concerns in this post); I am here to debate about whether search engines should use human behavior data to make search engine results more relevant, especially Google. Google analytics, Adsense, and other channels that Google can tap directly into, give Google plenty of data pertaining to users’ browsing behavior and search. If Google somehow feed this data (either directly or indirectly) in its search algorithm it could make search results more appropriate than what they already are, much closer to perfection. It would be as if the search results were handpicked by jury members themselves. For example if a page gets a large number of clicks for a particular keyword irrespective of its low ranking, it could very well be the case that people find that particular web page more relevant to their searched query than compared to pages occupying the top slots. If this evidence was somehow used in tandem with other parameters for ranking, the web page in question could be moved to one of the top slot and make SERPs more relevant. Other data such as bounce rate, average time spent, etc. could be used to make similar modifications.
Let’s face it. Search engine spiders have a long way to go till they become capable of replicating human behavior with a high degree of perfection. Since it is not feasible to handpick search results because of the shear size of Internet, search engines will have to find ways to incorporate human behavior into search engine algorithms to make it better – more human.
I am sure not everyone would agree but I would like to hear about you comments, nonetheless.
