Why customer churn prediction is what your retention team needs.


Grace Yang

If you listen to music often, I’m sure you know of the Apple Music vs. Spotify rivalry in the music streaming industry. And usually, you’re on one side or the other—but not both. In this multi-billion dollar industry, companies like Apple Music and Spotify are forced to regularly produce updates and keep their subscribers happy. Then, in 2016, some researchers looked into what makes each company successful, and they found that Apple Music had a churn rate that was three times higher than that of Spotify. But what exactly does that mean, and what made this information useful for the researchers at the time?

What is churn?

Let’s start off by defining what a churn rate is: churn is the rate at which customers stop doing business with a company over a specific period of time. In other words, churn measures how fast customers are leaving a company. In our music streaming example from above, churn is the rate at which customers stop subscribing to Apple Music, Spotify, or any other streaming service. More specifically, since Apple Music’s churn rate was much higher than Spotify’s churn rate in 2016, more of Apple Music’s customers were leaving than Spotify’s customers were. 

We can calculate churn by using the following formula:

churn rate for a specific period = (customers lost ÷ total customers at the start) × 100%

While it is possible to calculate a company’s churn rate, there is no predetermined “good” churn rate that all companies should aim for. Ideally, companies would have a churn rate of zero, which would mean that they would not be losing any customers at all. But of course, that is very unlikely, and almost impossible in the real world. A more realistic way to utilize and compare your churn rate is to compare it to your company’s previous churn rates and your competitors’ churn rates. 

So why should companies pay attention to their churn rates? The reason is tied back to the importance of customer loyalty. It is widely known that it is much cheaper and easier to retain current customers than it is to get new ones. In order to get new customers, your company has to do a lot of work marketing and reaching out to new potential customers, whereas to keep your previous customers, your company needs to send effective communication and stay connected with customer needs. Therefore, churn rate essentially has two effects on a company’s revenue: a high churn rate means that more customers are leaving, so you receive less revenue, but it also means that your company needs to spend more money obtaining new customers, which also hurts revenue. In other words, it is beneficial to your company if you decrease your churn rate. In fact, some experts say that decreasing your company’s churn rate by just 5% percent can increase your profits by up to 125%.

Now that we’ve established what churn is, let’s dive a little deeper into the different types of churn and what they mean. In particular, let’s look into the difference between soft churn and hard churn. We can define soft churn as the behavior that indicates churn, or just before a customer cancels their subscription or stops buying from a company. On the other hand, hard churn is when the customer actually unsubscribes or stops buying from the company. Let’s still consider our music streaming example from before. Let’s say that a Spotify subscriber regularly opens the Spotify app and plays music. But in the past two weeks, the user has not opened the app once. This behavior might indicate soft churn since they have not actually canceled their subscription yet. Then, two months later, the same customer cancels their subscription. This is hard churn. It seems like the soft churn predicted the hard churn, and this is what we will focus on in the rest of this blog post: how to predict churn and utilize it. 

What is Customer Churn Prediction?

Let’s define churn prediction as the process of identifying customers who are likely to churn, i.e. stop doing business with a company. Churn prediction can be incredibly helpful, which is why it is recommended to identify soft churn and customers most likely to hard churn and then attempt to reverse it before it’s too late. 

There are several things that churn prediction can help us do, including launching reactivation campaigns, allowing for proactive customer engagement, and creating targeted offers. 

Reactivation Campaigns

A reactivation campaign reaches out to, or “reactivates” customers who have stopped doing business after a predefined amount of time, and attempts to pull these customers back into the business. 

Let’s take a look below at an example of a reactivation campaign email from a fictional music streaming company, Streamify Music.

Subject: 🎵 Exclusive Comeback Offer Inside!

Hey Grace,

Miss the beats? We do too! Come back to Streamify Music and enjoy exclusive Free Premium Access for 3 Months! Offer valid until 02/02/2024.

Discover new tunes, old favorites, and curated playlists. Your music, your way!

Let's hit play together,

The Streamify Music Team

P.S. Your ears deserve it.

Reactivation campaigns, like the email above, are useful because they help retain customers and encourage more business activities, like buying. However, there are many details that need to be considered when it comes to implementing a reactivation campaign. First, it is important to establish how long the “grace” time period is for a customer (how long should the company wait before contacting a customer). This time period varies depending on the customer. Some customers may regularly interact with and buy from your company, while others may interact less regularly. Thus, it is important to calculate predetermined time periods for each customer in order to send them the appropriate number of emails/communications. It is also important to customize communication based on a customer’s “grace” time period and their buying habits. All of this is challenging, but can be made easier through the effective collection, storage, and analysis of customer data, which we will discuss later. 

Proactive Customer Engagement

While the purpose of reactivation campaigns ise to pull back customers and “fix” the situation, the purpose of proactive customer engagement is to anticipate customers’ needs and wants before customers reach out or before a problematic situation arises. Reactivation campaigns are reactive, while proactive customer engagement is, well, proactive. And the best way to be proactive is through constant and effective communication. 

Let’s take a look at an example of an email that aims to proactively communicate with customers. 

Subject: 🎵 Sneak Peek: Be the First to Jam with Our Latest Beats!

Hey Grace,

Hope you're groovin' to the rhythm! We've got something special brewing at Streamify Music, and we want you to be the first to experience it.

Unlock exclusive access to a sneak peek of our latest playlist feature:

Get Exclusive Preview Now

Your opinion rocks! Share your thoughts in a quick survey and snag a special discount on your next premium subscription:

Share Your Thoughts and Unlock Your Perk

Thanks for making Streamify your music hub!


The Streamify Music Team

P.S. Your beats, your influence – let's shape the future together! 🎶

In order to implement more proactive customer engagement, there are many things we need to keep in mind. First, you need to collect, store, and analyze customer data on preferences. You also need to maintain two-way communication with your customers by establishing easy ways for customers to provide feedback. 

Proactively communicating with customers helps make customers feel valued and more satisfied, but there are a few challenges we should mention. First, it is important to determine the optimal amount of communication a customer should receive to avoid making customers feel annoyed and avoid losing business at the same time. And of course, there is the challenge of having enough data and being able to utilize it. 

Targeted Offers

Another thing that churn prediction can be useful is through targeted offers, which are personalized offers for customers based on their preferences. 

Let’s take a look at the targeted offer email below for our customer, who especially loves country music.

Subject: 🤠 Exclusive Country Vibes Await You, Grace!

Hey Grace,

Hope you're riding the country waves! 🎸 At Streamify Music, we've noticed your love for country tunes, and we've got something special just for you.

Get early access to exclusive offers from your favorite country artists! Upgrade to our premium subscription now and enjoy:

  • Early access to new releases
  • Behind-the-scenes content
  • VIP invites to virtual meetups with country stars

Your country playlist deserves the VIP treatment, and we're here to make it happen. This exclusive offer is valid until 03/03/2024, ensuring you're at the forefront of the country music scene.

Have any country artists in mind? Hit reply and let us know – we might have a surprise in store for you!


The Streamify Music Team

As we can see, targeted offers are another great way to increase customer satisfaction, maintain customer loyalty, and increase revenues. 

More Churn Prediction

So, now we know what reactivation campaigns, proactive customer engagement, and targeted offers are and how they can be accomplished with churn prediction, let’s focus more on the details of their implementation and where churn prediction and Cotera come in. 

The first step is to collect and bring in all of your relevant customer data into a data warehouse, such as Snowflake or BigQuery. This data could be engagement data, order data, or feedback data, just to name a few.

Then, an in-house analyst should build a model to predict your company’s churn. As mentioned before, churn is an important measurement of a company’s success. Alternatively, apps like Cotera can help predict churn. See this case study for how we do it.

Finally, we should use a tool to sync our analysis and results into your company’s marketing and communication. In other words, this step is needed to take your data and results and then actually create the emails that we saw above. This step can be accomplished using reverse ETL systems, such as Census. 

While these steps may seem intimidating at first, they can be extremely helpful for a company’s revenues and success. Moreover, Cotera can actually help companies with all of these steps and improve customer satisfaction and retention. 


Throughout this blog post, we’ve been able to understand that churn is the rate at which customers stop doing business with a company. We’ve also seen that churn prediction is useful. In particular, churn prediction aims to prevent hard churn and maintain customer loyalty, and it can be applied to reactivation campaigns, proactive customer engagement, and targeted offers.

Clearly, churn prediction is beneficial for companies. Now it’s up to you to begin implementing these ideas in your own company. 

Let me tell you a story about customer experience ...

We run a podcast interviewing best in class CX teams. We get to know how they are helping their businesses win.

We'd like to share some stories of what good CX looks like, and what, uhhh! - less than good looks like. Are you in?