If you have read our article on how to identify the sources of churn then by now you should have everything you need to create a steady stream of feedback about all areas of your product. That means you are now ready to use that feedback to actively reduce churn.
But once you have a plan in action how do you know what the actual impact of your work is? How do you measure the effects of each change that you implement?
Sure, you could measure your churn rate each month to see if it is trending up or down. That’s simple enough. But your monthly churn takes into account all users on your platform, on different plans, who joined at different times and have different use cases.
With so many variables it muddies the water on how effective your changes can be. For a real-life example, let's say you make a change to your onboarding based on feedback from customers and your analytics. Over the next three months, your churn rate remains the same. Would you classify that change as successful?
Many would say it was not successful (in the worst case) or it was ineffective at reducing churn (best case). However, in our scenario, it's entirely possible that everyone who churned had no interaction with your onboarding change as they were existing users. In this example you can see how simply measuring overall churn can be ineffective as a method for assessing our efforts to reduce churn.
But there is a method we can use to accurately track our efforts using what is called cohort analysis.
What is Cohort Analysis?
A cohort analysis is an analytics technique that focuses on the behaviour of a group of people over time, allowing you to uncover insights about that specific group.
But what does that mean in the real world? Let's take an easy example: If we placed your users into a cohort analysis like the one below where they are grouped by the month they signed up and then measure their churn rate.
The analysis would show you based on the month of sign up how many churned each month. This allows you to map the churn of new users against the dates that changes were made to reduce churn.
If you made a change to the onboarding process to reduce churn in June and after three months everyone who signed up in July had a lower churn rate than those who signed up in January after the same period of time, you can clearly demonstrate that the changes you made had a positive effect.
By constantly referencing the date that changes were made to the platform and comparing the cohort analysis you can clearly see the impact that your efforts are having.
How do I create a cohort analysis?
To create a cohort analysis you will need software that can handle such functionality. Depending on how advanced your data visualization is you can use specialised software like Tableau that will automatically update each month. Or, you can use a tool like google sheets or excel that must be created manually each month.
Both will provide you with the visualisation to assess the impact of your changes.
If you already have a specialist tool like Tableau there is a strong possibility you have someone in your team who could create this for you if you explain what information you need to visualise.
However, if you do not have this person in your team, go to any of the popular freelancing platforms and you will be able to find someone who can create this on your behalf.