Why every business should be using topic modelling
When businesses receive large volumes of complaints, it's vital they harness that data for improvements.
Yet, how can businesses understand large amounts of fragmented data and turn it into meaningful insights?
This is where topic modelling comes into play
Topic modelling is a tool that sorts through text and categorises it into understandable topics. For example, sorting through customer reviews and complaints to understand common issues and how often they arise.
It is the most efficient way to sift through large volumes of feedback and review data. You couldn’t expect a business to go through each review one by one thousands of times.
At CourtCorrect, we’ve been exploring topic modelling and how it can benefit businesses when reviewing large amounts of complaints data. Let’s explore this further in this blog using publicly available review data about HSBC.
What can topic models show us?
We plotted public HSBC reviews over time and frequency and focused on the most recent, frequently discussed topics within the review data.
We found the most commonly raised HSBC reviews this year were about poor customer service and branch issues, such as phone wait times and branch closures, as well as credit card payment problems.
Isolating HSBC’s poor customer service reviews indicates a significant rise in disgruntled customers.
Using these models we isolated the individual reviews with the topic, to understand what led each individual customer to complain. For example, a large number of HSBC’s customer service issues arise because customers recently found their Direct Debits aren’t being set up correctly for their credit cards.
One of many reviews where customers are finding their Direct Debits aren’t taking their monthly payments.
Each review offers specific feedback that businesses can use to gauge customer experiences over time. Additionally, it helps businesses prevent new complaints from arising in the first place by fixing root causes.
Not just the problems
Because HSBC’s reviews are primarily 1-star reviews, the most common topics are about problems customers have experienced. But it is also important to identify the positive reviews, too!
We isolated the most common 5-star review topics to see what was going well, so that the business could advertise its strengths on and build out its competitive advantage.
We found that most of these positive reviews referred to great experiences with individual staff members whom the customers named. These reviews show that there were particular employees that went above and beyond for customers, and who can be learnt from as examples of how to provide a great customer experience. Additionally, focusing on the importance of the human touch may be a winning strategy to advertise and win new customers.
One of a cluster of positive reviews about a particular staff member.
What are you waiting for?
Customers are freely giving feedback on your business's performance. Neglecting to harness this information can leave you trailing behind.
The topic models we've created will give you access to these insights which could be the key to improving services, averting common problems and recognising stand-outs.
Failing to adopt topic modelling isn't just overlooking an analytical tool; it's missing out on a trove of actionable business intelligence.
Intrigued by the power of topic modelling?
Contact CourtCorrect here to gain access to your business’s topic models.