Last week the New York Times uncovered a story about an online vendor who has managed to leverage his bad customer service reputation into a top ranking on Google’s search engine.
The business owner was quoted saying “I’ve exploited this opportunity because it works. No matter where they post their negative comments, it helps my return on investment. So I decided, why not use that negativity to my advantage?” This strategy appears to have worked. The more furious the online chatter, the higher the site ranked in Google search, resulting in greater awareness of the brand and an increase in sales.
Google blogged saying “that being bad is, and hopefully will always be, bad for business in Google’s search results. And after quickly putting together a team Google is satisfied that they have developed a solution and it is already live.
Google outlined some obvious ways of solving the issue including blocking the offender, using sentiment analysis and placing user reviews next to search results, therefore exposing the negativity towards them. Instead they have created a new algorithm that detects extremely poor user experience and have incorporated it into their search ranking as an initial solution.
In this particular case, incorporating sentiment analysis in search ranking results is the right thing: shady business no longer profits from bad customer experience. But is sentiment analysis always a good thing?
There are reasons why we might not want Google to determine good sentiment vs bad for us in its web ranking. For example, if we wanted to look up something political and there was negative sentiment surrounding a particular leader, it may prove very difficult to learn more about a political situation. Sentiment analysis may also filter out bad reviews that are often very useful in doing research when making a purchase. The internet has always been a great resource for finding both good and bad reviews and information.
Although it seems as if Google has squashed the idea that any publicity is good publicity, there are sure to be more complex search discrepancies coming our way. Sentiment analysis has not been, and most likely will never be, completely mastered. The intricacies and nuances in language are difficult enough to comprehend face-to-face, let alone trying to develop an algorithm to decipher and make sense of it all (think: sarcasm). So, for now, we should continue to use our own intuition when considering sentiment analysis in search and maybe even dig a little further if we can’t find what we are looking for. After all, we can’t rely on Google to do everything for us.