In the last 7 years, especially after Google’s first attempt at semantic (meaningful), search called, “Knowledge Graph” was launched, there’s been a forward movement on this kind of search.
Semantic search tries to understand a searcher’s motive through contextual meaning. Intent - determination, desire - is a human emotion. A computer was unable to understand it because machines were not programmed to comprehend human emotions; the tech for that was not mainstream. Till a few years ago when a conscious attempt was initiated to try understanding search words in the context they were being made.
Semantic search got impetus because of the advent of digital voice assistants which sparked off voice search. Now, as most of you would know, spoken and written language are never the same so the words used by a voice “searcher” can differ from those used by his “text” counterpart, even though they may be looking for the same thing.
Here’s an example: If an online searcher keys in: I need to change the oil, the person could be referring to: (1) car engine oil (2) cooking oil (3) two-wheeler oil (4) snow plow, etc. Earlier, a search engine would use the exact search terms to throw up the results. Today, however, it will try and understand the searcher’s intent before throwing up the “relevant” content around the query. That starts from establishing who the “I” is? Is it just an ordinary person or someone qualified like a mechanic? Is it someone located in Canada or India (the chances of the query being related to a snow plow is more if it is a Canadian rather than an Indian is raising the issue)? If intent is not accurately captured it means a blow to all three parties involved - the searcher, the content provider as well as the search engine. The searcher, for he did not get the relevant answer, the content provider, for his content was thrown up against the wrong query, and the search engine, for failing to help the searcher.
Click here to read the rest.