Sentiment Analysis: The Misunderstood Metric

Sentiment Analysis: The Misunderstood Metric

If your job title contains any variation of the words “marketing,” “community,” or “brand,” you have very likely heard of sentiment analysis. Also known as opinion mining, this is the process of combing through a large amount of post-level social data to determine an audience’s feelings about a certain topic.

Here’s how it works: once a social audience is identified, all of their posts relating to the topic are pulled in to an audience intelligence platform and analyzed. The posts are then classified into one of three categories – positive, negative, or neutral. It’s frequently used to gauge an audience’s response to a new product design or marketing approach.

People Pattern - Audience Intelligence - Sentiment Analysis Buckets

Sentiment analysis has been used in a surprisingly diverse range of applications, from planning messaging strategy in presidential campaigns to car sales to travel planning, and the technique has received quite a bit of attention in both the marketing community and the mainstream press. The industry response to sentiment analysis hasn’t been entirely positive, though. A few jaded marketers have called the technique a fad or – gasp – a vanity metric. Even on this esteemed blog, we once accused it of “lacking nuance.” Which is true… in some cases.

The truth is, while analysis isn’t a magical all-in-one solution, it is a valuable part of the larger audience intelligence toolkit. Below, I’ll walk you through a use case where the technique really shines: monitoring an audience’s response to a new policy.

Example: Online School

It’s an understatement to say things are a bit bizarre in the U.S. right now. In response to the COVID-19 outbreak, many schools are moving to an online-only model in an effort to prevent further spread by minimizing close contact between students. So, our question is this: how do students feel about the new paradigm of online school?

To start, I pulled 9,555 Twitter profiles that mentioned online school in the past seven days into the People Pattern platform. I filtered down to individuals located in the U.S. who explicitly identified themselves as students. Then, I looked at the posts, filtering out irrelevant positive matches (i.e. people talking about online college programs such as University of Phoenix). I also discarded retweets and posts with neutral sentiment.

Students’ Response

People Pattern - Audience Intelligence - Student Sentiment for Online School

So far, students are responding extremely negatively to online school. The value in sentiment analysis here is telling us there’s a problem. In this case, the problem is clearly statistically significant and severe enough to invest our time into investigating.

Looking deeper into the post-level data, the primary objections seem to be difficulty focusing at home, teachers communicating poorly through online learning platforms, and technical difficulties with videoconferencing apps. Surprisingly, lack of in-person interaction with peers was not a commonly cited concern for students. 

In the positive column? Students love ordering UberEats for lunch.

Think of sentiment analysis as a monitoring tool for audience response. It can quickly let you know whether your audience likes or dislikes a new campaign or policy change. If the results show an issue, you can delve deeper with other tools offered by audience intelligence platforms, such as keyword/phrase analysis or simply going through posts the platform has tagged with a negative sentiment to diagnose the problem. From there, you can tailor your campaign or make appropriate adjustments to the policy.

Interested in learning more about how an audience intelligence platform can address your specific needs? Click here to schedule a demo with our team.