Emoji Analytics: A Social Listening How-To
Graphical iconography expresses what words can’t – and in a more economical manner. The language of emojis and emoticons has exploded and brands are beginning to discover emoji analytics to find the sentiment of messaging around their brand. Emojis have become part of our daily lives. When you’re rushing through the market between appointments and someone sends you good news, there’s no time to respond, ‘That’s amazing! I’m so happy for you.’ Instead, you just plop in a smiley face, hit send, and you’re on your way.
Businesses have largely ignored emojis – aside from some standard-bearers like Pepsi and Starbucks, who have created customized iconography for their brand advocates to share. But how do you track the analytics (volume and sentiment) around graphical imagery when social listening is built on a foundation of text analysis? As technology has progressed in synch with Moore’s Law, parsing and analyzing content on both the open and closed web has become faster and more efficient. Throw in a couple of algorithms and some machine learning and voila – your social listening dashboards are a cocktail of influencers, mentions – and emojis.
It’s important to understand the impact emojis and emoticons can have on a brand’s Social Listening analysis. The addition of a graphical layer of communication can turn a seemingly innocuous message into an expression of pure vitriol with the inclusion of several angry faces. An Instagram post which contains a brand-image tagged solely with an angel-face emoji immediately transforms from a neutral mention into positive sentiment.
Below are two business use cases which show the importance of emoji/emoticon detection and analysis to any company’s Social Listening program.
Sarcasm has long been a problem in machine analysis of human language. The subtext and tone of snark, innuendo, and irony can transform a seemingly positive message into one of bitter negativity. Take Instagram post above as an example. When analyzed through a purely text lens, the post seems overwhelmingly positive. However, if you’re a brand manager at Atlantis and you investigate into a bit more detail, you’ll find that the posted photo was far from #stunning (image recognition has even further to go than language analysis) – and that the inclusion of two angry emojis flips good vibes to bad. A vertical that relies on good word of mouth and positive customer feedback to drive revenue simply cannot ignore the sentiment expressed by their patrons. Without emoji and emoticon detection and analysis, a sentiment-driven business misses daily opportunities to address problems with customer service.
The Apple App Store and Google Play are home to a combined 4 million apps. Approximately 25% of those apps are games. App developers who don’t have the marketing dollars of heavyweights like King and Supercell are often resigned to casting their efforts into the sea of a million apps. Imagine the number of sleepless nights – praying for a positive review in the app store, because reviews often translate to more prominent positioning within the game’s category. Those nights might continue to be sleepless because gamers tend to congregate in forums – and that’s where they share their feelings about new releases. Social listening can help to aggregate and organize the chatter on forums and review sites about a new mobile game. In doing so, patterns will begin to emerge from resonant channels and key influencers. Here’s where emoticon detection and analysis is so important. The ultimate goal for the insomniac app developer is to pinpoint users with positive sentiment and subsequently activate an outreach program, whereby happy customers are spurred to return to app stores to leave glowing reviews. In this case, every user counts. Sleepy McCoder can’t afford to miss even a single brand advocate. In the example below, TheBlueBomber is a powerful influencer on the popular video game forum, Neogaf. He’s expressed his intent to download – and his excitement. Pure text analysis doesn’t record excitement. It records a colon and a capital D. Emoticon detection and analysis is smart enough to recognize the relative positioning of the two characters and relate to them a positive value for sentiment. Our developer’s business depends upon accurately translating the confusing vernacular of his consumer-base, so it’s critical that his listening software is smart enough to know that 😀 means ‘I’m excited.’ With emoticon detection and analysis, each influencer’s dialogue is appropriately translated and categorized for sentiment. Relationships are solidified, reviews start appearing on ‘official channels’ and Sleepy is sleepy no more.
If our developer isn’t using emoticon detection and he’s filtered his dashboard to show only positive mentions, he might miss his opportunity to connect with TheBlueBomber and nurture the relationship.
As you probably know by now, Synthesio recently announced our emoji analytics and emoticon analysis feature within our Social Listening platform, so if you want to learn more about this, and see how it works for yourself, then reach out and request a demo to see it in person!