How Big Social Data Meets Creative Marketing: Sentiment Analysis

This week, Felicia Panosoglu from the global team gives us her take on big social and creative marketing.

Let’s face it- most businesses that try to keep up with the pace of the changing digital ecosystem fall into the “gather all” trap. Aimlessly collecting data and tracking customers without actually thinking about the point of doing so. As a result they overlook some of the most valuable insights they can find out about their audiences though big social data — sentiments.

In the early noughties, Web 2.0 ushered a new era for the online world which saw user-generated content peaking alongside dynamic web pages. So came the unheralded arrival of micro-blogging and posts offering opinions about everything and anything (including brands), anytime and anywhere. Basically we wanted to tell the world how we felt about things, love them or hate them, we said it.

This influx of user generated posts created a new arena of “opinion mining” for marketers to enter. At the bare bones, what it allowed us to do was get a reaction point from people towards an issue, through the analysis of a given data set. This became known as sentiment analysis. At its core was linguistic analytics, which identified positive, negative or neutral sentiments in a text. Simple right?

Sentiment analysis became a tool used widely by marketers to define strategies based on the emotional tone of consumers. This has since evolved with sentiment analysis shifting from being a mere tool for insights, to becoming the primary material of marketing. In short, it’s not only being used for marketing, but also in marketing, moving from insight to activation.

Just like we care about each other’s feelings, brands also need to change how they speak to people according to their emotions.

Sentiment analysis for marketing

Sentiment analysis provides an incomparably wider scope of data compared to traditional marketing analytics like focus groups and surveys. Social media monitoring is sentiment analysis. For example, Twitter users ubiquitously express their opinions on different topics, products and brands. Research shows that 20% of Tweets that mentioned a brand expressed a sentiment or an opinion concerning a product or service.

Sentiments are something all brands should consider as a tool to measure their brand awareness and consideration levels. But that’s only half the story — you’re only measuring sentiment after a campaign has run. The added value comes in using sentiment analysis to inform the way you activate your campaign.

Sentiment analysis IN marketing

Instead sentiments can be used to influence how a brand talks to you in real-time. Coca-cola’s brand’s slogans around happiness show how closely the brand is already intertwined with sentiments. But they went a step further in their Coca-cola Happiness Machine campaign using sentiment analysis for marketing.

For the Coke Happiness Machine campaign, the brand identified a tiny little town in Spain as the least happy city in the country. They were able to identity this through natural language processing among Tweets and their location.

Coca-cola decided to do something about this. They began with placing Coke vending machines across the city. To get a Coke, customers had to give their Twitter usernames to the machines. The machines then instantly conducted a sentiment analysis to the customers’ tweets, and gave them a happiness score. If the score was high enough (ie. If you were happy enough) then you could get a free Coke, if not, you paid the price determined by your happiness score.

It was a new approach to dynamic/real-time pricing and the internet of things. The machine conducts the analysis in real-time and determines the price that the customer will pay instantly, allowing for flexible pricing.

The effect didn’t just generate chatter but stayed true to their brand strategy by their consumers ‘choosing happiness’ through buying coke.

Thinking beyond likes and shares

For brands, the effectiveness of a strategy or a campaign should not only be measured with preconceived metrics like ROI, but also by cutting-edge models of evaluation.

Of course, human sentiments go beyond negative, positive or neutral, and there are advances in sentiment analysis that dive deeper into wider sentiments. The one thing we can be sure of is that humans have sentiments, and they go beyond likes and shares. By using sentiment analysis intelligibly, you can not only understand your customers better, but also touch their feelings.