In this series, we take a closer look at some analytics projects for our energy customers. In a previous article, we already discussed an innovative customer journey tool. Once again, we will let Jan Vercammen speak. He will tell us more about a Churn insights tool, which is helping to solve one of the biggest issues plaguing the utilities sector.
Because utilities companies are operating in a saturated market, few of their customers are on their first energy supplier. Churn and acquisition are therefore two important terms within the sector. Utility companies have a tendency to focus on holding on to their clients rather than on bringing in new ones, because the former strategy is much cheaper by far. In other words, a thorough investigation into the reasons why customers want to leave a company is no luxury.
High-tech model predicts churns
It is, of course, difficult to predict when and where a customer will choose to take their business to the competition. AE consultant Jan Vercammen explains:
“Utilities don’t play a major role in the lives of their customers. They are rarely aware of who their supplier is, until a problem presents itself. A company can only start from vague signals from their customers: has the customer, for instance, contacted the help desk?”
Based on those signals, it is of course difficult for utility companies to see a churn coming. This is why AE co-developed an advanced churn model wih an utilities customer. The tool predicts the behaviour of utility company customers. Obviously, the model can never guarantee 100 percent accuracy. However, it does provide useful indications that a customer will most likely change their energy supplier.
This is how the Churn insights Tool works
When we survey the customers of a utility company, we eventually obtain a complex model. This model consists of thousands of smaller models. To make such a complex model understandable and actionable, AE used cutting-edge machine learning explainability techniques to uncover the inner workings of the model. This enabled us to explore individual customers’s churn risk and risk factors over time. These insights were made accessible to business users via a Churn Insights tool that we co-created.The high-tech tooling uses internal data, but also takes open data into account to analyse a customer’s complete environment.
“We use the full data base of all events that were compiled with the Customer Journey tool. This way, we can predict when people will decide to take their business to the competition and we also gain insight into the reasons why,” Jan continues.
A concrete example
The tool revealed that the age of customers is an important predictive factor of a churn. Young people clearly show different behaviour than people in their forties. The behaviour of the latter group in turn, deviates from the oldest segment. These insights can be explained on the basis of the lifestyle of each age segment.
The ultimate tool for successful marketing campaigns
The tool has even more potential. It is namely possible to select segments and add different variables. Jan Vercammen tells us more about focussing on a specific group of customers:
“We can, for instance, focus on young people who have been customer for some time and received a payment reminder at some point in the past year. It then becomes very easy to set up a concrete marketing campaign that is specifically geared to this nicely defined target group. That is a major advantage of an internal data science team and shows that the Churn insights tool is also ideal for marketing purposes.”
The Churn insights Tool spurs to action
The co-created end result provides utility companies with an excellent view of their customer portfolio. Thanks to the calculated models, they get a clear image of customers who are about to churn. They can then quite easily work out a strong business case and take timely action when customers want to switch to a competitor.