Okay, so you know the drill:
You just bought a product. Or read a blog post. Or saw a video. Etc and etc. There's usually a link to share your review on the best thing ever/horrible thing whose creator deserves to be flogged. And thanks to smartphones, people are no longer tethered to computers to shoot off an opinion. You can go buy something, open it up, decide it sucks, skewer the item and return it before you even leave the store.
Nearly everything can be commented on by just about everyone online. With all that information floating around, you just knew that someone was going to start compiling it all, didn't you?
By someone, I mean computer scientists, and how they?re compiling it is called sentiment analysis. It's an emerging field that attempts to translate human emotions ? in all its warm, fuzzy or furious glory ? into data. Sophisticated algorithms can identify more than just general feelings about specific topics; they can even pinpoint influential opinion holders.
No doubt businesses are extremely interested in this. These days, online consumer opinion can make or break a product, and these companies know it. Case in point: I have a friend whose complaints about a certain nameless car dealership was only addressed once he bought the domain name www.XYZAutoStinks.com, and forwarded it to the president of the company. Suddenly, they bent over backwards to resolve his issue.
What's the takeaway here? If you?re getting a raw deal from an otherwise reputable company, don't forget you have the power of your opinion to scare the heck out of them. (Just make sure to deal with someone high-ranking enough to help you. A rank-and-file service rep probably won't cut it.)
On an individual level, there are already basic sentiment analysis tools for Twitter data, like Tweetfeel, Twendz and Twitrratr. According to Tweetfeel, 77 percent of recent Tweeters felt positively about the film ?Julie & Julia.? (Well, that's kind of neat to know. I was interested in it anyway, but now I really want to see it.)
Twitrratr, on the other hand, put a negative score to the comment ?julie and julia was truly delightful!!? because it ended with ?we all felt very hungry afterwards.? In this case, the word ?hungry? was mistaken as an indicator of a negative sentiment. So neat or no, sentiment analysis is not a perfect science yet. But it has great potential to become a powerful tool, especially as its systems and algorithms get more refine over time.
Beyond Twitter, I?m looking forward to what this science will do for search. Now, if you plug in 'the best restaurant in Philly,? you get deluged with every page that has ever included those words, relevant or not. Instead, I?ll someday be able to get actual restaurant, hotel or movie results ranked by users? opinions.
There are 30,500,000 results for 'the best restaurant in Philly.? Are you kidding?
So it turns out Mother was wrong. If you don't have something nice to say about something, go ahead and put in there. Keep sharing those thoughts, good or bad, about the coolest phone or dumbest app. The most verbal among you are influencing our whole collective online consciousness.
[via The New York Times]