Canceled Flights to Insights

At Infegy, we love traveling and looking for great deals when exploring (ask us about our flight perk). We'll use this love to highlight how Frontier Airlines could use diverse social listening datasets to better respond to customer complaints. We're not doing this to single them out, but because people often express their complaints about airlines online. Being stuck in an airport with internet access and nothing to do often leads to venting customers. This makes airline customer complaints a fascinating dataset to analyze.

To analyze this phenomenon, we'll compare data collected from the Frontier Airlines’ subreddit with Infegy's general social listening data. These are very different datasets, serve different purposes, but inform the same conclusions. Infegy's out-of-the-box social listening data has a much larger volume but is more general in conversation topics around airlines. On the other hand, the subreddit data is much more targeted but has a lower volume, consisting of Redditors expressing their negative experiences with Frontier Airlines. By applying suitable filters, both datasets inform each other and can provide valuable insights on how Frontier Airlines can address customer complaints effectively.

Why we're using Frontier as our example

Social media users frequently criticize budget airlines like Frontier Airlines or Spirit online. They attract travelers with low-cost fares, but there's a catch. These airlines charge extra fees for many extra services like snacks, change fees, or additional baggage. Moreover, their gate agents earn commissions by catching customers with oversized bags, resulting in hefty fees for checking those bags. As a result, social media, especially TikTok, has been flooded with posts about these practices, and many of them have gone viral with hundreds of millions of views. These emotional social media posts make an excellent target for analysis.

Comparing Topics across datasets

We’ll begin our analysis by comparing how people discuss Frontier Airlines' baggage fees in two places: a general social search and the Frontier Airlines subreddit. At first glance, these discussions seem different. First, Infegy's social dataset (including news stories) has 36 times higher volume than the more niche subreddit dataset. Second, the sentiments expressed in both datasets are dramatically different. In the general social dataset, 54% of the discussions about baggage fees are positive, whereas, in the subreddit, 66% of posts are negative.

The volume differences have an obvious reason. Infegy's collection engine grabs social posts at scale from all corners of the internet. Conversely, Frontier Airlines' subreddit only has 2,200 subscribers and only a few hundred comments over the last few months. The reason behind the sentimental differences is less apparent. We get our first clue from the Topics themselves: the social dataset's subjects are more elevated and appear more professional. At the same time, the subreddit's issues are minor and often express anger.

Why do the conversations look so different?

We'll look at a two-year channel distribution of Frontier Airlines fee conversation to dive into our hunch on the linguistic differences. Starting in August 2022 (when Infegy added news content), News makes up almost 60% of the entire Frontier Airlines conversation. While this is helpful for analysts researching Frontier's press coverage about a brand, some researchers care more about genuine, organic conversations. So, to achieve that, we'll apply a channel filter.

Excluding News to reveal the organic conversation

We'll use a channel filter that excludes news sources like the New York Times or the Wall Street Journal to focus on authentic conversations. You'll notice that both datasets now appear more balanced. Our social dataset volume decreased by about 80%, indicating that people might discuss Frontier Airlines less than viral TikTok videos suggest. This drop in volume is excellent news if you're an analyst at Frontier: a significant use case of social listening data is determining whether viral moments threaten your brand or are more bark than bite.

On the other hand, sentiment also normalized. After removing News, Infegy's social dataset sentiment dropped by 23%. The Topics in the social dataset now sound more conversational and less formal, suggesting that we are visualizing more genuine, organic conversations.

Frontier's subreddit vs. filtered social

Now that we know we're getting more organic conversations, we'll dive into the clusters of conversations themselves. We'll do this using Infegy Atlas's Narratives, which compares cross-document topics at scale to uncover commonalities across social media posts.

To read the entire insight brief, visit our website!