When you perform a search on Google, two major things happen within a fraction of a nanosecond. Google parses and processes your query in its index and attempts to find a closest match or answer to your question. Second, Google tries to guess or predict the “subject” and offers answers that Google thinks you are searching for. This depends on a lot of factors which may include your previous search queries, list of visited sites, social context, interests, behavior and so forth.
Consider the following example search query:
Google thinks that I am searching for “miss”, which is a traditional honorific for unmarried women. Additionally, Google thinks that there is a chance I am searching for an album by John Mitchell. The fact however, is completely different from what Google thinks I am searching for. I am searching for the verb “miss” which means “to fail” as in – he misses the target and turns back to the crowd.
Now you may be wondering whether Google precisely understands the words you use in your query. The short answer is No it doesn’t. Google and other search engines have no clue about the meaning of words used in the query and they don’t rely on the meaning alone for fetching results. However, what is important is the context of the query, whether the words used in a query have something in common or whether they imply a question whose answer (closest match) is already available in the index. When you search for “tear”, Google can only “predict” whether you are looking for a waterly fluid or the ripping apart of something by force.
Here is another example search which shows how crowd sourcing affects Google search results to a great extent.
When I typed “make money” in Google search box, Google assumes that I am someone who wants to make money online. There are a thousands ways of making money, but Google thinks (and predicts) that this guy is looking for online money making ideas. In reality, I am looking for the novel by Terry Pratchett in the Discworld series. Unless I use the identifier “novel” in this context, Google will assume something based on what millions of other users are searching for and show the same results to me.
Then we have a SPY world to consider which introduces a social mix in regular Google search results, showing you relevant content and information from your social circles. But not everyone is using Google Plus, which at this moment is nothing but a virtual ghost town used by only Google employees (and of course Robert Scoble).
Hence the Plan B.
In coming months, Google is all set to radically change the search mechanism with a novelty called “Semantic search”. According to a report from Wall Street Journal, Google will soon implement its semantic search technology to the search filter, which will analyze the meaning of words typed into the search box. Based on the meaning and context of words and phrases, Google will provide answers directly on the SERP page, instead of throwing a list of webpages, maps, ads or links.
The current keyword and Pagerank based ranking system isn’t going anywhere but I believe we are entering an age where results are not ranked merely by Pagerank authority. Google fellow Amit Singhal told WSJ that semantic search will process information by understanding the actual meaning of words. An excerpt from the post:
Google search will look more like “how humans understand the world”. For many searches today, we cross our fingers and hope there’s a web page out there with the answer. Some major changes will show up in the coming months, Google is undergoing a years-long process to enter the “next generation of search.”
Mind you, Google already provides answers to simple questions e.g the population of India, the capital of Australia and the author of The Tempest, so there is nothing new here. But according to the WSJ report, semantic search will tightly integrate with Google search, offering answers and close suggestions that are contextually relevant.
This can mean a lot of things, other than just plain answers. If someone performs a search for a famous singer or a band, Google might show all the music albums and DVD’s of the current artist on the search result page itself. This can be disastrous for sites who rely only on the long tail of search, buy tons of links for a specific keyword and lure users to their affiliate commission sales page. For if users find all the music albums of a particular artist directly on Google.com, they surely won’t click the fourth result on the SERP page.
On another note, this can be the false front for a forthcoming war on Apple’s Siri, remember Google Majel? Eric Schimdt has already declared Siri as a competitive threat to Google’s mobile search so this move completes the equation.
There is another guy with the same label of “an answer search engine”. Bing.
Well done Google, I remember they copied your results way back.
The ability to learn faster than your competitors may be only sustainable competitive advantage. ~ how true!