Real-time bidding on emotions - a possible future with sentiment analysis
New York, New York - the city of heatwaves, flash floods and electrical storms
It has been a few weeks since I landed in Stockholm, greeted by the cooling weather instead of the heatwave provided 35 degrees of New York City. I had the privilege of experiencing an all inclusive Mother nature package with electrical storms, flash floods and heatwaves. The latter was what I experienced most of the time and, after a few days, I did actually start to enjoy the weather, from inside my apartment, with the AC on max. I’ve been in NYC to attend a conference and learn more about the upcoming trend within the digital world; the sentiment analysis.
From Donald Trump to kitchen blenders
Nowadays we spend hours in front of our desktop or mobile device. The majority in Sweden use these devices for: social media, habitually shopping online and, first and foremost, looking up information. This data is scattered over 1 billion websites and contains useful material in regard to different enqueries. Whether it’s looking up information concerning the presidential campaign, where for now Donald Trump is the most searched for Presidential candidate, or which kitchen blender you should choose, we turn to the internet.
Rational choice theory + the internet = tears of joy
Rational choice theory has many aspects, one of which is the trait of people being well informed when making decisions. Thanks to the internet and all of its information, we are getting one step closer to reaching tears of joy from the makers of rational choice theory.
Turning to the internet eases the decision making process when, for example, buying a new product or service. By looking up information about the product or service, on different websites and social media, we can take facts and opinions into the decision making process and thus, hopefully, take better decisions.
The information we base our decisions upon usually contain positive and negative opinions towards a brand, product or service and always have one thing in common; they are written in text or expressed through images.
As an organization, offering a product or service, wouldn’t it be wonderful to regularly tap into this kind of information and learn about people’s opinion regarding your brand or offer? For instance, what if you could harvest this information, process it and offer an even better product or service? This is where the sentiment analysis comes in. In order to find out the emotional state of a text, image or other sources, we need to analyze it, this is what the sentiment analysis is all about.
The sentiment analysis could be performed manually by a person going through text and images noting the emotional state of each, for example, if a person is expressing; joy, disgust, irritation or any other sentiment. However, this could take a great deal of time and the more information you have the longer it would take. Worst comes to worst, the analysis is done after too much time has passed and there is nothing you can do about the expressed sentiments regarding your product or offer. In order for this not to happen, the ongoing trend is to automate this process, that is, instead of people we use machines to analyze the sentiments.
Almost there, but not quite
Judging by what I have seen so far, we are not quite there when it comes to a fully functional automatization of the sentiment analysis. However, I do believe it’s just a matter of time before product launches, movie releases, debates or any other event could analyze the real-time emotions of its crowd. For example, which political opponent had the majority of positive sentiments expressed towards them and vice versa during a debate. Another example would be when releasing a trailer or movie at the cinema, did the crowd react the way we expected them to during the happy/sad/horror scene?
Disgust, anger, sadness, excitement, joy
One of the most gripping instances I experienced during my stay was the example to the left. I stood in front of a camera which analyzed and stated the emotions I expressed with my face. It was accurate most of the time. Imagine having these kind of cameras on, for example, mannequins or test audiences. You could get immediate feedback regarding which product people enjoy and disliked the most.
The camera also worked when there were more than one person in front of it. Now imagine you are a company wanting to run an ad at a sports event, without the sentiment analysis you would probably buy a time slot and your ad would show at that specific time no matter the circumstances and sentiments, be they good or bad. In this case, the risk of associating your product with a negative sentiment can be quite high and you thereby jeopardize having paid a generous amount of money only to attain negative attitudes towards your own product, this is not ideal.
Instead of associating your product with random sentiments, wouldn’t it be great if you could make sure that your product is associated with a certain, not random, type of feeling? For instance, envision having the aforementioned cameras at the event and that you could run the ad when 90 % of the crowd are expressing joy, wouldn’t that be amazing? This is what sentiment analysis can bring to the table and perhaps, when the sentiment analysis has reached its full potential, it will be a game changer to the marketing and other industries. Perhaps the time slots of different events will function as a real-time auction, just like the online marketing world, where instead of bidding on a search term you bid on an emotion and hopefully win the most valued time slot, when most people express the desired sentiment.
How soon is now?
In some time, this will most likely be our reality. We will be able to make real-time decisions based on real-time sentiments regarding an event, the latest product line or a brand etc. The question is only when.
Any thoughts? Feel free to drop a line if you have any thoughts whatsoever, I'll gladly discuss them with you!