12,000 Angry Men – Measuring anger within high volumes of customer complaints
When people are angry they want to be heard, and nowhere is this more apparent in the digital world than on social media. In the same way that social media has provided companies with a new channel for promotion and advertising, it has also provided consumers with a new outlet to talk about their experiences and, in many cases, with a much louder voice.
“A happy customer tells their friend, an unhappy customer tells the world” as the well-known saying goes.
At Precise, we are heavily involved in analysing customer experiences.
Recently a client of ours, with their big toe deeply submerged in the ever evolving telecommunications ocean, had a business need to understand pain points in customer experiences for key providers to develop potential solutions to help mitigate them. Their brief was simple but clear: “we know people have negative experiences, but we want to know how negative they feel, and what really triggers their anger”.
By the very nature of the medium, with posts being written at such pace and quantity, it is easy for organisations to feel overwhelmed by the volume of customer feedback sent to their customer service Twitter handles, and not be able to analyse data to identify the root causes of complaints.
We therefore needed to find a way to delve into the content and build a framework to identify the pain points that cause the most anger to the majority of customers.
We therefore developed a metric which could help capture different levels of anger. We decided to classify Twitter posts on a scale from ‘unhappy’ to ‘irate’ with several gradients in between. To identify ‘irate’ posts, we took on board Mark Twain’s suggestion and only qualified a post as ‘irate’ if customers appeared completely exasperated with the service, expressed a loss of all trust, used bad language consistently, or threatened to move provider.
We based our analysis on Twitter posts from customers to the handles of five major telecommunication suppliers within the UK. The study incorporated quantitative and qualitative research, with the latter being based on the analysis of random samples of posts to gauge sentiment, levels of anger, and the reason for complaints.
This allowed us to provide a detailed view of customer pain points across telecoms providers as a whole, and a benchmark for each provider, identifying the worst and least bad performers. We backed-up our findings with numerous verbatim, many, of course, unsuitable for family viewing.
The findings proved that our simple yet creative approach to measurement in this instance had enabled us to effectively tackle the client’s specific question and make sense of large volumes of data.
As a result of our conclusions the client was able to prioritise actions and equip brand community managers (and customer service representatives) with additional knowledge to assist with the delivery of a more satisfying customer experience.
Sean Farrell, Brand Analyst, Precise,