Reducing churn - insights from a churn analysis

Sound familiar? If your organisation offers subscription services, reducing churn is probably on your agenda. It may not be the most popular topic: we naturally prefer to focus on what is going well. But what is the best way to analyse churn? Where do you start? In this blog, we explain how you can use data to analyse churn, with the hope of reducing it. 

What is churn? And why is reducing it so important?

Churn is the number of customers who stop using your company’s services or products. It shows how well a company retains its customers. Churn has a direct negative impact on your organisation’s revenue and profitability. Losing existing customers means an immediate drop in recurring revenue, and acquiring new customers is often more expensive than retaining the ones you have. 

And what about your competitive advantage? Retaining customers makes your revenue more predictable, which brings stability. It lets you focus on growth rather than replacing lost customers. This can help you stay ahead of your competitors. Besides, what could be better than a strong product and satisfied customers?

1. Collecting and organising subscriber data over a given period

In our analysis, we focused on subscribers’ online use of two news websites and their associated apps.

Our goal sounds simple: improve subscriber retention. We want to understand why some subscribers cancel their subscriptions and identify the factors that contribute to this. By gaining insight into these patterns, we can develop targeted strategies to improve subscriber retention.

For our churn analysis, we started by organising a large amount of previously collected subscriber data. The data was spread across several tables, so our first task was to examine those tables carefully and bring the relevant information together. We selected the most important factors based on domain knowledge and supported by a literature review (Belchior, L. M., António, N., & Fernandes, E. (2024).

To combine the data, we used dbt (Data Build Tool) together with Snowflake. Snowflake serves as the central repository for all our tables, while dbt lets us transform the data efficiently and systematically. With dbt, we can build models that select the most important columns, perform calculations and ultimately create a combined dataset. This allows us to process our data consistently and repeatably, giving us a robust foundation for further analysis.

After consulting domain experts, we chose an analysis period of one year. This period provides a good overview of subscriber behaviour. One drawback is that seasonal factors were not taken into account in the analysis. This could affect how we interpret the results, particularly if churn patterns fluctuate considerably between seasons.

Because our dataset is so large, it is essential to split the transformations into short periods. In this case, we chose to add data to our dataset monthly, although the right interval depends heavily on the number of customers you work with. We used the Conveyor tool for this task.

2. Defining churn

An important step in analysing churn is establishing a precise definition of it. The definition you use can have a considerable effect on the analysis, so it makes sense to base it on the analysis’s purpose. Because we are interested in the behaviour behind churn, we chose fairly strict rules. For example, we did not label subscribers who returned within two months as churners. This helps us avoid noise in the dataset, as many subscribers temporarily cancel their subscriptions to get an extra discount. If you are more interested in revenue or monthly subscriber losses, it may make sense to include these subscribers in the analysis.

3. Data analysis and interpretation

After combining and organising the data, we identified the key variables that could affect subscriber churn. We then analysed these variables to find patterns indicating a higher risk of churn. During this analysis, it became clear that grouping subscribers into segments was essential to refine the insights.

We chose to base the segmentation on three main criteria: subscription type, the length of the subscriber’s relationship with the organisation, and subscriber activity. These segments allowed us to see which factors played a significant role in churn risk within each group. 

By segmenting the data in this way, we could better identify patterns and behaviours specific to each subscriber group.

Conclusion

Analysing and addressing churn provides the following insights:

• Churn analysis shows why customers leave, helping you improve your product.
• Identifying churn risk factors enables proactive retention strategies.
• Reducing churn pushes you to keep adding value for your customers.
• By actively tackling churn, your subscription business can retain its customer base, grow and improve its services. This leads to a stronger market position and sustainable long-term success.

Do you also want to understand the value of your customers and segments? Read how we approach this through customer analysis with data.