Imagine if you could capture and make sense of the data that measured the global or regional mood. If you had the tools to measure whether the mood is upbeat, like the feel-good factor created by the recent Olympic Games, or maybe the downcast mood on bad economic news coming out of the US or Europe.
That kind of valuable information could be used by marketeers to pitch tailored products at opportune moments targeting when consumers are more likely to be in the mood to buy. Similarly, such nuanced and detailed indicators could be used to improve user experiences and build brand loyalty through exemplary customer care.
Sentiment analysis is nothing new - we just have new ways of capturing multiple digital sources and making better sense of the data. Derwent Capital, a hedge fund, used sentiment data harvested from aggregated social-media feeds like Twitter to return 1.86% in its sterling shares over one month. That level of return beat the market and was ahead of most other hedge funds at the time. In the context of multibillion-pound funds that’s a significant return on equity.
According to IBM, we create 2.5 quintillion bytes of data every day - that’s a lot of potential data that can be refined into useful information. That data might come from any intelligent device that can gather and transmit in some way. As our world becomes more connected and woven together by the "internet of things”, there will be an increasing amount of potentially useful data that is interpretable. Making sense of that data is a growth sector as the refined information is such a rich source of understanding for creating tailored marketing directed at individual consumers. People are often brand loyal - if you can win them and then support them supporting your brand using sentiment technology you have created a sustainable virtuous circle.
The potential uses of sentiment technology cross all sectors and geographic locations. From data collected from African farmers using mobile technology to check market prices for grains and the resulting buy and sell data, to user satisfaction at entertainment venues. Sentiment analysis brings a new emotionally intelligent level of potentially deep insight into what, how and why people are using and consuming. Mobile devices add further depth to that picture of user-generated content by piecing together user habits as we live our lives wherever that might be.
Case study: Sentiment analysis at work
The mapping technology that underpins the mobile internet model is integral to building sophisticated pictures of usage - taking data and information about you and applying it to wherever you are and making suggestions about what you might like in that specific location. It could be anything from suggesting that a particular local restaurant is good because your mobile technology knows that you have a liking for a style of food. Or it could be that your technology knows you are a cycling enthusiast and there happens to be a renowned cycling shop a hundred yards away.
Whatever your take on sentiment technology, whether you feel it’s intrusive or useful, or maybe a bit of both, it's established and getting cleverer and more orchestrated. There are nearly six billion mobile phone subscribers globally. Google’s one million or so servers alone deal with an estimated 24 petabytes of data per day. As smartphones become the default and big data gets bigger, there will be a lot more sentiment gathering in our increasingly electronic landscape.
Steven Hipwell (pictured) is principal project manager at Birmingham City University
This was first published in October 2012