When marketers analyze free-form text for insight, such as reviews or social media posts, they often see the same sequence of approaches, which become increasingly sophisticated along the way:

  1. Descriptive statistics like word count, sometimes represented in word clouds, or frequency of keywords over time
  2. Sentiment analysis: tracking the mix of positive, negative, and neutral posts
  3. Emotional analysis: tracking the mix of posts by the presence of the “fundamental human emotions” – made famous by the Pixar move Inside Out (although there is disagreement in the scientific community just how many there are)
  4. Topic modeling: identifying the various topics of conversation that exist through natural language processing (NLP), especially in a large collection of posts (i.e. segmenting the conversations by topic)

This toolkit is still very useful for analyzing words. However, the “word analysis” approach becomes problematic because people are expressing themselves less with words on social media platforms.

Enter emojis. Some are common (👍), while some may be more familiar to emoji experts (🔥). Others are just weird (💩)!

Thanks to sources like emojipedia.org, we can get an objective translation for most emojis. But how can we put them to use as marketers? Emojis may mean fewer words to analyze, but we believe they may provide more information.

For example, algorithms have difficulty identifying whether a comment is sarcastic based on the words alone. So do humans! Some emojis, however, such as the eyeroll (🙄) or upside-down smiley (🙃) often self-identify a comment as sarcastic. Brands who consider these emojis in their analysis could help customer service teams, community managers and strategists identify and resolve issues.

Another area where emojis could help brands is by better understanding the emotion being expressed, or how intense it is (there is a great deal of interesting work being done in academia in this area: Exhibit A and Exhibit B, for example). Instead of merely positive sentiment, brands can use emojis to create a scale of positive sentiment, like these tweet replies from two satisfied Panera customers who are about to receive free coffee:

Holding other things equal, the tweet without the emojis is positive, but arguably not as much as the other, and the difference is worth measuring.

Marketers have long known that strong advocacy, not the ambivalent kind, is tied to bottom line growth. A Net Promoter Score (NPS) only counts ratings of 9 or 10 as promoters, after all. If a respondent rates your brand as a 7 or 8 it is not even part of the NPS formula. We should similarly strive to define scales for positive and negative sentiment with the help of emojis, and measure progress in moving consumers along that scale.

Emoji features are not waiting for marketers to catch up. Multiple emojis can be combined to create a message (✈️😎). Increased variety, adjustable colors, avatars, and current event-inspired trends are (😷)  added dimensions we need to deal with. Not to mention differences in interpretation across cultures. From a data scientist’s perspective, developing the algorithms to generate insight, and reliable automation of those to make use of insights at scale, will be difficult. From a marketing strategist’s perspective, the use cases are endless. But since emojis have become part of the vernacular, and are arguably a new and free data source, they cannot be ignored.