Everyone knows about the awesome power of the Google ‘Empire’ and the omnipotence of its search engine algorithm, but is it so clever that it can actually differentiate between positivity and negativity?
We already know that the Google search engine is capable of applying something called ‘stylometry,’ which was originally used to determine the authorship of anonymous writings or classical music.
What it amounts to is studying the style in which a piece was written, to determine if it follows a particular pattern, although the methodology used to do this is quite complex.
Googles Methods
Google’s use of stylometry is along similar lines and means that its search engine algorithm is able to determine the usage of a particular word, even when it has more than one meaning. A good example of this would be the word “date” which can mean a fruit, an entry on a calendar or a social meeting.
Some words can even be used as both a noun and a verb to make it even more confusing.
Disambiguation
The method used is known as ‘disambiguation,’ which is the process of defining the intended meaning of a word i.e. removing ambiguity. To do this the search engine has to show a level of artificial intelligence by understanding both the user’s query, as well as the context in which it is being used. This means that it must analyse not only the word being used, but also the style of writing around it, to determine the context.
Sentiment Analysis
Assuming that this is the current level of ‘intelligence,’ this leads us on to whether Google can determine enough about context to tell the difference between a positive review and a negative one, for instance. Evidence suggests that it can – Google filed a patent in 2007 titled “Domain-Specific Sentiment Classification” and began using it in July 2011.
Quote from Patent No. US 7,987,188 B2:
“For example the word ‘small’ usually indicates positive sentiment when describing a portable electronic device but can indicate negative sentiment when used to describe the size of a portion served by a restaurant.”
For this type of differentiation to be effective, Google’s algorithm would have to consider the adjectives used in a piece of text to make the analysis of positive, negative or neutral, which is also a part of the patent.
At this stage, nobody is entirely sure how far advanced Google is with the development of tools for ‘sentiment analysis.’ Supposition is currently that it is already being used for local search, Google Places and product results. How long it will be before it is used on a global scale remains to be seen. It is almost certain to be a part of Google’s algorithm updates in 2012.
The Effect On SEO
For the SEO community, it means that we must encourage our clients to employ excellent customer service methods, nurture their communities and ensure that every possible step is taken to avoid getting any negative reviews in the first place. Building positive relationships with clients requires a good deal of time and effort, probably more than most businesses care to give, but the way things are going, it is rapidly becoming an essential thing for any E-commerce website designer.
Image Credit: David Castillo Dominici
Quite informative article. Google is the king of search world. I am surprised that Google can define the intended meaning of a word. Example of word “date” is very true. Great advice to establish good relationship with clients.
Thanks Mark.
I think that Google is just thinking for the sake of all people who use their engine. But, I must say that it can be hard for SEOs to do work without Google.
Google has really played its part in the world of internet. So, I guess everyone should be happy about it. Even though, sometimes Google’s algorithms are not that accurate(that’s what I think).
interesting post thank you for sharing this post, i like reading anything to do with Google i find it very interesting.
Hi Mark,
This is really an interesting post about Google and their SEO methods, apparently Google is trying to make their Search engine but on a experience or interest based results which means than the same results will never be the same for two different people using the same laptop. Thanks for sharing.