A Holistic Approach to Understanding Churn
2020 Winner

A Holistic Approach to Understanding Churn

Bonnier AB

Sundsvall, Sweden

Category Best Use of Data Analytics or Research

Overview of this campaign

Mittmedia needs to tackle churn if we don’t want to run out of customers within a predictable future. Often there is not one cause for why customers churn and often there is not only one action that can be done to counter churn. In order to take effective action, Mittmedia needs to communicate and understand churn from a holistic perspective. Only if all parts of the organisation see how they can contribute to reducing churn without cannibalising on other departments’ actions and effects, then we will be able to reach a long-term success in reducing churn. We gathered representatives from customer relations, the editorial staff, but also from our CTO for a brainstorming session on how churn insights would be used: 

  • First, a churn initiative would be used for operative business development: How to target customers with a high risk to churn? What are these customers’ characteristics and how do we retain them? From the editorial side, the operative question was more from a content perspective: which content do we need to produce to satisfy high-churn risk readers so that the risk for churn diminishes?

  • Second, churn insights would be used on a more strategic level. Knowing why customers churn and which kind of customers stay with us, business management would be able to have more information on the features that our end customers appreciate in our products. It could also give a hint on which customer groups we should invest most in for the future. Our editorial could get more insights into what topics to cover to retain our current customer groups, but also what we need to write more of to gain new readers.

  • Third, the insights from a churn model should result in churn predictions in the long run, so that the company can move from being reactive to proactive in the long run. Those predictions would be crude at the beginning and evolve in time as the churn model evolves and gets better.

  • Fourth we would need to find the churn factors which we can directly act on vs churn factors which we can adapt to. For instance, the age of our readers is not something we have in our control, but knowing which age groups don’t read from us, will help us in forming strategies forward. 

We decided to take account of those four needs when starting the first iteration of churn models.

 


Results for this campaign

The visualization of our results can be seen in detail here:

https://docs.google.com/presentation/d/1q3eIyBPwq9zR4zk9xNisyKqaJagZFvyh6w1Nrer5ehg/edit?usp=sharing 

Our churn model gave us the following insights and resulted in the following actions within the organisation:

  • The number of push notifications which a reader gets and opens correlates highly with churn. Customers who have received at least 20 notifications during a month tend to cancel their subscriptions earlier than those who have received fewer. This resulted in an editorial call to action to revise the way pushes are done. It has been decided to redefine news-value for articles in order to use this for a new - and maybe automated - push logic.

  • Women churn unproportionally more than men. Already in the first 100 days we loose 10% more women than men by churn. An editorial initiative has been started to understand what different genders read and how we present gender in texts. It turns out that we have a skewed reader group where women are underrepresented in texts and choice of topic in our articles and that we write more articles of interest for men. Our churn results spurred actions to write more balanced.

  • The age-groups 25 to 45 are churning the most amongst both genders for our readers. This is also the age-group which has most money to spend. In this group we have the strongest potential in increasing subscription length. Our department for editorial business are now considering new approaches for targeting those age-groups.
  • We compared the users who are 90% likely to churn to the users who are only 10% likely to churn and can see that users with a higher likelihood seem to be in general less active in our products, since they almost don’t watch videos at all, they often come indirectly from social media to our products, they are not interacting with the disabling function for push notifications, they don’t use airplay so often and they experience more average payment errors. We are now looking into causality and are setting up hypothesis test for engagement. 
  • We should concentrate on the customer group which is 80-60% likely to churn since we have the best chances to engage them in a way to stay with us and we should take action during the first 3 months since this is the period where most churn occurs. Our communication and customer department is devising strategies considering our churn results. 

 

We have visualized and communicated the churn results in several presentations to all departments and it has spurred many actions for counteracting churn. You can read much more in this blogpost: https://medium.com/mittmedia/we-looked-into-why-our-subscribers-churned-with-machine-learning-d1ac1adf800


Contact

To contact a company representative about this campaign, click here for the INMA Member Directory