It is highly important that all companies within any sector focus much attention on dealing with churn. Because customer churn can easily end up harming the business revenue numbers and through influence management decisions.
Customer churn rate refers to the percentage of consumers that leave your company during a given month, quarter, or year. Some examples of customer churn for SaaS firms include the closure of an account, non-renewal of a contract or service agreement, cancelation of a subscription, or using another service provider.
So way do churn occur? Well, to be honest, there are many different reasons why consumers can choose to leave your company. A churn analysis can help to identify the causes of churn. This will give you a better understanding and consequently provide you with opportunities to implement effective retention strategies.
If you already know your customer churn rate but need knowledge and help to reduce it, you can read our article about “18 Simple Ways to Reduce Customer Churn“.
In order to make the most beneficial decisions in relation to retention strategies, we advise you to dig a bit deeper. Do not just settle on practicing traditional retroactive churn management that is based on a simple statistical model. Rather, you should be implementing a risk analysis-decision making segmentation approach. This kind of analysis will be an enormous help when trying to retain your users.
So what exactly is a predictive churn model?
Well, it is a classification tool for your consumers. By looking at your consumers’ activity from the past and examine who is active after a specific time you can design a model that classifies the steps and stages when a consumer is leaving your firm.
With such a predictive churn model, you will gain great knowledge and quantifiable metrics that you can use in your retention tactics. The model identifies patterns of habits among the customers who leave. Then you can step in and take action before your users make the final decision of churning.
If you do not have this kind of classification tool, chances are you would be operating on general assumptions, rather than on a data-driven model that indicates your consumers’ true behaviors.
Retaining your buyers without a strong understanding of their behaviors is highly difficult. Therefore, the first step in building the predictive churn model is to understand your consumers’ action from customer data points.
Customer churn is normally divided into two different categories; voluntary or involuntary.
You will be able to get a broad understanding of how and why your users cancel their accounts, by analyzing how these two types of churn influence your company.
Once you have that information, it is much easier to develop a strategy to proactively prevent churn.
Voluntary churn is when consumers actively choose to cancel their subscription. This cancellation can be due to a number of reasons, such as:
It can be hard to hinder users from voluntary churn because the underlying reason is usually complicated to solve. Yet, it is possible to try to prevent this type of churn. For example, if customers are considering leaving your firm for a competitor, they might be persuaded to stay if you remind them of the value your service can give them.
Involuntary churn is when a customer’s account is canceled unintentionally. This can happen because of the following reasons:
It is often easier to prevent involuntary churn rather than voluntary churn. This is due to the fact that involuntary churn is almost always caused by the fault of a mechanical failure. In many cases, this can be avoided by implementing a proper system of dunning emails and notifications. Alternatively, you can apply an automated tool such as ProfitWell Retain, to help reduce churn due to payment failures.
Once you understand the difference between voluntary and involuntary churn it can help you segment your customer base. This will allow you to build a more extensive churn model.
As mentioned earlier it is hard to retain your consumers without a strong understanding of their behaviors. So when creating a predictive churn model you first need to gain an understanding of your consumers’ actions from customer data points.
Below you can find a list of the kind of data you need to collect, in order to examine the triggers that causes your buyers to leave your business.
All of the above data are just some examples to get you started. Obviously, the more you know about your users and what event leads them to leave, the better. A great amount of data will give you a more accurate model.
Besides collecting a great amount of data, you also need to start asking the right questions so you can discover “triggers” that lead to customer churn. At this stage of creating your churn model the most vital part, is to ask as many questions as possible. By doing so, you can test, retest, and qualify your hypotheses and data before you start to implement a model.
You will see that most of the questions have two core variables. These are quantity and time. Below you will find examples of three questions you can use, to get an understanding of how to start your process of data investigation.
1. What is the percentage of correlation between available data points and churn?
This type of question could give you answer such as: “73% of all our churning customers had registered an online complaint”, or “68% of our lost customers have never downloaded the free software”.
By getting these kinds of answers, you would hopefully be able to see a pattern of correlation in particular from data points to declining users.
Additionally, you need to make sure you ask the same questions to your current customers. Then you will know that you aren’t just making assumptions about something being a direct cause of users leaving. Chances are that the same percentage of your current users share the same data values.
2. At what point in the product lifecycle did your consumers leave?
At what stage did your customers leave?
Was it when the subscription expired and they did not renew it? Was it related to a holiday or other external events?
How long did your users stay before churning? Did they stop to use your service before they officially left?
Getting answers to all of these questions is vital to understand the critical points in the consumer journey and where you need to focus your attention.
3.What was the login pattern? Or how many times did they use “X”?
By looking into your users’ login patterns and frequency of usage, you will receive more detailed data over time. Here you are searching for something unique about the pattern and quantity that isn’t related to a current subscriber’s pattern and quantity.
In many instances, you will start to see a few high correlation attributes. Finding those special correlation triggers is essential to the final step in the model creation.
Putting it to Together
Once you have asked all the right questions and received all the relevant data, you will start to see some interesting patterns of churn.
It is possible that you will find some very high correlated data points that are not as differentiated as you would have hoped. We recommend that you find a least three high correlation attributes. If you can’t – keep digging!
In those cases, you are probably not asking the right questions or you do not have the right data points in your collections just yet.
When, at some point, you have all the right churn triggers identified you are ready to start designing the final churn model for your business.
By now, you should have a system in place that can distribute specific churn scores to your customers. Following, you have to start the next phase, which involves testing and learning into your response program.
The predictive scores you will get from the model will be assigned to your users on an individual level. Therefore, you need to create a methodical approach where you can test the most efficient response system.
Some of the key elements you need to test includes:
The wonderful thing about this type of testing is that you will get a clear indicator of success: an enhanced churn rate.
As you have probably discovered by now, creating a predictive churn model is a comprehensive process. It takes a lot of creativity, patience, and strategic thinking. During the process of building the model, you will collect some amazingly rich and actionable insights. Make sure to use this information wisely!
In the end, you will have created a response system for your business that will keep more of your users coming back. This will be at a measurable level, so you also have reliable information at hand.
We, therefore, urge you to get started on creating your own predictive churn model for your SaaS business. Because when your revenue is based on recurring monthly or annual subscriptions, every user that leaves your firm creates a hole in your cash flow. A high customer retention rate is necessary for your survival.
If you need any further assistant on how to run a churn predictive modeling process, you can watch an on-demand webinar about predicting customer churn.
We are confident that you are now ready to start building your own prediction model for your SaaS company.