Churn analysis: predicting customer churn [Updated 2023]

Einblick Content Team - September 29th, 2022

What is churn analytics?

Churn analytics is the process of measuring and understanding the rate at which customers quit the product, site, or service. Churn analytics can help you understand how frequently customers churn out of the product and where this tends to occur. It can also help you understand which features and functionality are important for keeping customers in your product. Churn analytics is critical for getting a performance overview, identifying improvements and understanding which channels are driving the most value. You need to include churn analytics in your agile development framework or product roadmap work plan, so that you can take actiochn based on your findings.

Churn analytics can be tracked in many forms – by customer, by week, by quarter, and so on. Research shows that a good or “acceptable” churn rate for a product or service to be successful is between 2% and 8%. There are many factors that determine the ideal rate of churn including the product’s offering, competition, and sales cycle duration.

Effectively producing customer churn analytics can be challenging as many stakeholders, from the product team all the way up to the CEO, stand to benefit from a deeper understanding of customer behavior. Having a good tool to help you analyze customer churn data and ultimately predict customer churn, so you can prevent it, is essential for any business.

At Einblick, we provide cutting-edge data science software in the form of a uniquely visual and collaborative canvas. While in Einblick, you can connect to datasets and databases, code in SQL or Python, build machine learning models using AutoML with our progressive computation engine, and more. Due to our progression computation engine and cells, which package commonly re-used snippets of code, you can vastly accelerate your time to insight. In addition, you can collaborate live with various stakeholders, sharing your work, thought processes, and progress quickly and effectively. This collaborative platform is ideal for customer churn analysis and predicting customer churn as you can work with stakeholders of all technical levels in the same workspace. In an increasingly remote-first work environment, truly collaborative software is needed in every industry, including data science.

Check out our canvas demonstrating churn analytics on bank customer data:

Read on to learn more about churn analytics and how you can predict customer churn.

What should I look at when analyzing churn?

Total churn rate: The total number of customers who have left the product, site, or service. Their reason for leaving could have been poor product quality, poor service adequacy or any other factor. You should look at the entire picture: for example, do you honestly think that a particular feature may have had an effect on churn? It is important to look at these factors in detail and not make assumptions without doing further research.

Customer churn rate: This is the number of customers that stop using your product or service within a given timeframe. Churn rate is important because it helps you gain a clearer understanding of your monthly recurring revenue (MRR). Thus, churn analytics are an important tool that can help you to make informed decisions regarding your business.

Short-term vs. long-term churn: Those who leave after a short period will be classified as short-term churn. Those who leave after a long period can be classified as long-term churn. For example, a short-term churned customer may drop off during a free trial period, whereas someone who doesn’t stop using the product until 5 years of using the product, would be considered a long-term churned customer. Given that different factors may be influencing short-term vs. long-term churn, you may consider analyzing the data differently based on the kind of churn you are particularly interested in.

Returning visitor rate: This represents the number or percentage of the visitors who return to use of the product, site, or service. Some analytics tools will set an expiration date on visitors to be considered “new.” For example, let’s say that visitors to your site are considered “new” for 2 years. If a “new” visitor comes back to the site within 2 years, then they are considered in your returning visitor rate because they have come back to the product within the set timeframe. Active users are much less likely to churn than one-time or infrequent visitors to the app.

Referrals: It is normal for high-value active users to refer to people in their network who in turn may or may not convert into active users. You should factor this into your churn analysis by tracking referrals in your data warehouse. For example, if a user referred five users and only three of them converted, what is its impact on future revenue from the referred users? To quantify such impact, you should consider: if these “referred users” were all new customers, what would be the revenue of each of these new customers? If they were all existing customers, how many of these customers would likely refer to others? These are just some of the issues that need to be considered before you start tracking referrals. Accounting for the entirety of a customer’s lifetime value requires careful consideration. Churn analytics will allow you to answer this question: what’s the lifetime value of a new customer?

Churn metricsChurn metrics

How to measure churn rate

The way in which you measure the churn rate of a product hinges on your key performance indicators (KPIs). If your goals include the number of months a user has subscribed, you’ll need to know every time they unsubscribed.

If you are measuring churn specifically against time, and you have a way to track when people unsubscribed, that connection gets slightly more complicated. The key is that you must maintain measured consistency across time. Once you choose a metric, you have to stick with it. For example, you could always measure a current user as having been a monthly active customer on the day they submit a payment for subscription, until the day they cancel their account(s).

If you offer some kind of subscription service where customers get charged regularly–let’s say every month or every year–make sure you have a way to measure customers’ time as active users. Additionally, if you are accepting payment through a third-party system, that system may release funds or offer payouts every month or 6 weeks. During that time, there could be some customers that are renewing their subscription on a different timeframe. Your team of data analysts then need to disaggregate the data from each payout to understand what each customer did during that time frame. This seems like common sense, but we’ve seen analysts miss something crucial here based on someone forgetting that each payout period counts as one month towards time metrics, while recurring subscriptions could be on a different time scale.

Here are some suggestions and considerations as you move towards a more robust system for analyzing customer churn:

  1. Determine how long each existing customer has been using your product.
  2. Decide how long it’s reasonable for someone to be an active subscriber before moving them to a different part of your statistics database.
  3. Analyze what percentage of people who subscribed for your service or product continue to do so. They could be automatically renewed, so perhaps they are subscribed but not using the product. Or perhaps, they actively renewed their subscription after the end of a free trial.
Churn prediction modelsChurn prediction models

Churn prediction models

Much of what we’ve discussed so far pertains to understanding existing customers or former customers, and why they may stop using a product or service, as well as factors that may complicate the analysis. But the goal for any business is to be proactive. How can you use churn analytics to make next month’s business better? You could, for example, determine how the last three months of customer churn data and analytics can help you make accurate predictions about future customer churn and monthly revenue.

Why do we need churn analytics?

Churn analytics can provide a snapshot of the value chain each month, showing how many customers are coming back to the product, whether customers are renewing their subscriptions, how much revenue each customer is driving, and how expensive each customer is in terms of acquisition.

Churn analytics can also help you to identify the causes of churn and determine ways to reduce churn. For example, is there a part of the product or service that is particularly cumbersome or tedious for the customer? How can the product team then pivot to meet those customer needs to lower the churn rate?

Why is it difficult to predict churn?

Churn isn’t always straightforward to calculate. The answer depends on the question that is being asked. If the question is: “How many customers will leave our account in a given period?” The answer is relatively straightforward. For example, if 5% of your customers leave every month, and you have 100,000 customers, then you would expect that 5,000 customers would leave.

If your churn calculation only includes customers that have paid subscriptions for a certain period, or you had a free promotion during that time, then it can become much more difficult to predict. Why? Because some customers are staying just for the free promotion, or perhaps there are external factors, like their company switches products.

Voluntary churn happens when a customer chooses to stop using a product or service–they actively cancel their subscription or purchase plan, or maybe they go from a higher tier to a lower tier. This is easier to isolate as a specific event in your logs. But what about users that are passively subscribing without using your product or service? When do we classify these users as having quit? There is also involuntary churn, in the case that a payment does not go through due to expired or stolen credit cards, or perhaps life events outside of your control. Churn analysis is useful for analyzing this behavior to understand why customers are staying in a certain product for shorter or longer periods of time than expected for their market segment.

Predicting churnPredicting churn

Customer churn analysis

Reducing the number of churn events and reducing the number of customers that leave the product each month should be your primary objectives when measuring and predicting churn rate. You will want to know what’s causing customers to drop off, when you can, and you will want to make this information actionable (so you can take steps to reduce customer churn rate).

How should we measure B2B SaaS Churn Analytics?

Churn analytics can be measured by a variety of metrics, including the number of users that renew their subscriptions, the number of users who cancel their subscriptions, and the average duration of a customer’s trial. Understanding and analyzing SaaS churn allows you to identify the causes for which your company loses high-value customers. SaaS companies should measure SaaS churn analytics as part of their growth plan, as it provides important feedback on how to improve the product.

Other Churn Analysis Factors to Consider

  • What needs to be improved about the product’s monetization policy?
  • Which of the customers in your sales funnel are actually using your product?
  • What is the value they are adding?

Value of SaaS Churn Analysis in a Business

With proper data insights, any team can take action to execute meaningful changes. As an example, the customer success team can identify unacceptable churn rates that are costing their company money. The product team can get a clearer understanding of where in the user experience that customers are having trouble. New releases or features can be made to alleviate these pain points. All of these actions lead to revenue growth and substantial business improvement.

Two types of product differentiation

Churn analysis is an important tool in product management to understand how a product or service is affecting customer behavior, and to improve the solution. When thinking about churn analytics and predicting churn, it is important to consider product differentiation–why are your customers using your product in the first place? It is important to distinguish between strategic product differentiation and tactical product differentiation when discussing what drives product upgrade conversion rate (and performance). Strategic product differentiation is more big-picture, in terms of understanding the market, your competitors, and what gives you an edge. Tactical product differentiation is more specific, what are the detailed actions you can take to make your strategic goals attainable?

In theory, all product offerings should lean heavily on one or more strong competitive advantages that create value for both customers and business units. But this can be challenging given that some decision makers may prioritize products that only generate sufficient customer value, without considering products that are highly differentiated. Understanding customer churn can help and work symbiotically with product differentiation.

For example, you want to consider:

  • How can we get more customers to trial the product?
  • What is a good pricing scheme for new customers?
  • Where should I place new pricing to get more value from existing customers?
  • Is there any seasonality in your recurring revenue streams? How much does the product need to be improved to support new lines of business?
Predicting customer churnPredicting customer churn

Customer Churn Analysis Basics

To calculate your customer churn rate and turn that information into meaningful results, you need to know where customers came from and how they got to your company. And in order to do that, you need to know how much value you, or your competitors, are providing to your customers and how effectively you are meeting the needs of these customers.

Nurture vs. Harvest: Customer Churn and the Customer Value Model

The customer value model (CVM) is the key to understanding how much customer loyalty there is and where it comes from. A CVM essentially measures the value that your company does or will bring to your customers. Customer churn analysis is one way of better understanding how effectively you are managing the customer experience from start to finish with every customer interaction. Then you can see if you are nurturing customer value, or showing them the value your company brings to their solution. The CVM helps you understand whether you are “nurturing your customers” or simply “harvesting them for your competitors.”

Customer Churn versus Product Features

What keeps churn rates low? It’s important to understand that just because a product is popular, doesn’t mean that it will decrease churn by itself. Often, if something is a smashing success, it will increase customers but not reduce churn. In other words, popularity has little to do with improved customer retention. A bad product could increase churn but a good product that is responsive to customer needs will almost by definition decrease churn. By showing customers that you see their difficulties or desires for the product by having good customer service and launching new product features, you will hopefully retain more customers.

The problem of churn statistics

It is important to remember that churn is only one part of an overall customer life cycle. As such it is an important part but it is not the entire picture. There are other factors in a customer’s life cycle including: customer satisfaction (CSAT), customer loyalty, and customer acquisition cost (CAC). Customer loyalty mainly reflects the long-term relationship with a product or service provider. Whereas CSAT and CAC may reflect more short-term efforts on the part of your company.

Proactive retention with churn analytics

To prevent churn, teams must be proactive about it. Teams can use a churn analytics tool to track retention at the individual level to pinpoint high-risk customers. Here are some options to try to reduce your customer churn rate:

  1. Use data to identify high-risk customers: To identify high-risk customers, teams can use a churn analytics tool or survey to survey their current customers. The tool will map out who has churned and identified their reasons for leaving. You can use the data to build machine learning models or other analytics tools to help you stay ahead of any customers at risk of churning based on prior data on similar customers.
  2. Add incentive to stay in the product, tool, or service: people stop using products every day. It’s natural for some people to leave because they don’t care about the product, but other people leave because they can’t pay their bills. Churn can be avoided if the company provides an incentive for people to stick around, even just for a few more weeks, this will make retention significantly easier to achieve. For example, a company offers a promotional code after a customer has not been actively using the product or service for a certain amount of time. Alternatively, you can also reward loyal customers by offering additional perks for those users who stick around (such as discounts or coupons).

How to optimize and reduce customer churn

Many of the reasons customers leave can be rectified before a user hits “cancel,” and a roadmap for a customer’s life cycle that includes customer success, engagement and retention is key to reducing churn rates.

To reduce churn, companies must focus more on the journey of customers. A good customer journey map should include:

  • Customer success and engagement: start by understanding each customer’s problem and from there identify a personalized solution to help solve their company’s problem. Save time by reducing the number of reports and questionnaires. Increase customer satisfaction through one-on-one customer engagement with a care team member.
  • Customer onboarding and adoption: make onboarding and adoption easier through training, onboarding automation and standardization of workflows. Set up a goal with concrete actionable metrics of targets set for each customer to increase their likelihood in using the product.
  • Customer relevancy: in order to get more out of your customer, you must make it easy for the right customers to find your product, stay relevant and build a clear brand. Be sure to have an FAQ page up with answers to the questions that are made frequently by your customers such as “What if I don’t like this?” or “How much do I have to pay?” Add animations to help customers navigate your site and make purchasing products easier.
Churn predictionChurn prediction

Preventing churn is everyone’s job

Having an engaged and passionate customer base is vital to any company’s success, but it is also important for your customers’ success as well. If you’re an online retailer, for example, you want to get feedback from your customers as soon as possible after an order is placed. Why? If you fail to address problems or concerns before they’re posted online, customers will leave negative reviews or new products and services may fail to attract interest.

Preventing churn is the partial responsibility of every person and every team within the company. It means maintaining a unified pulse, understanding what works and what doesn’t, and being proactive in anticipating problems. You should always have employees who are both motivated and invested in their work. This can be achieved by paying them better, motivating them with positive feedback, showing appreciation, giving them responsibility, and treating them with respect. However, building a team is a long-term effort that can take years. This means you should always have succession plans in place and never rely solely on one hire to keep your business running smoothly.

Churn vs. Retention

Although understanding customer churn rate and implementing churn analytics are really helpful, you also have to consider other metrics, like retention. The opposite of customer churn is customer retention—a business’s ability to keep its customers and continue to generate revenue from them. Make sure that you really know who’s using your product, who your customers are, what’s keeping the loyal customers coming back, what are your loyal customers’ use-cases, what are the tools that they use to complete the work that they bought your product for, how often do they use your product, etc.?

And ultimately, if you’re selling a suite of software and services, then it’s hard to ignore all those things that go with that. You have to be measuring them and if a customer is not happy with your product, then what do you need to do to make them happy? Are you still competitive in the market? Are you delivering value for money? You can use surveys and personal outreach to get insights into what’s driving cohort members towards or away from you so you can take proactive steps to retain them.

Benefits of Analyzing Customer Churn

Why devote effort to customer churn analysis when you could spend the time on a long list of other projects? Here’s a snapshot of benefits that accrue from analyzing customer churn.

  • An understanding of the lifecycle stage of customers can be a critical factor in policy development, management and distribution. The companies that derive value from customer experience and provide excellent service are the ones that provide effective customer management tools and develop effective relationships with customers.
  • Understanding the patterns of customer experience allows managers to target those team members who are most likely to provide an exceptional customer experience and not just respond to their needs.
  • Identifies new customers who require additional resources and sales competence. Identifying new customers who require enhanced knowledge and communication skills for optimal sales success.
  • Identifying churn processes and customers who are in the process of a churn cycle gives managers a heads-up that there could be additional resources needed for sales growth in specific markets or geographic regions. The resulting information can then be analyzed and used to enhance sales performance.
  • Customer churn analysis can help identify profitable customers, who may be in different stages of their renewal cycle and who, as a result, offer different opportunities for renewal. Sales management personnel do not need to spend their time on activities that do not bring value to the firm. Investing effort in revenue stream analysis will create focused opportunities for revenue generation.
  • Assesses risk factors that may have an impact on revenue performance. Segments with higher churn rates or low renewal rates may reflect greater risk factors driving their renewal cycles faster than other segments. Managers can focus efforts in those segments to identify alternative methodologies for improving customer satisfaction, or possibly inducing those customers to switch to other products or services. Such initiatives may be especially important for companies that want to expand into new markets, where competitors’ products or services provide a better overall solution than yours.
  • Provides insight into financial performance and profitability over time. Customer churn analysis gives managers a snapshot of companies’ financial results over time, including:
    • How a company is performing against its competitors
    • The relationship between financial performance and customer lifetime value
    • The correlation between financial results and customer renewal rates versus market growth rates over time.

Frequently asked questions

Customer churn forecasting is the process of predicting whether or not a customer is likely to stop using a product or service.

Churn forecasting is important because it can help a company identify at-risk customers and take steps to prevent them from leaving. By predicting churn, a company can focus its resources on keeping its most valuable customers.

Customer churn is typically defined as a customer who stops using a product or service within a certain period of time. For example, a company might define customer churn as anyone who cancels their subscription within the first month.

Common causes of customer churn include poor customer service, high prices, and poor product quality. Churn can also be caused by competition, changes in customer needs, and other factors.

Churn can be prevented through a number of strategies, including providing excellent customer service, offering competitive prices, and regularly improving product quality. In addition, companies can use data from churn forecasting models to target at-risk customers with retention campaigns.

Customer churn can have a significant impact on a company's bottom line. For example, if a company has a monthly subscription service with 100 new customers every month with a 5% monthly churn rate, it will need to acquire 125 new customers each month just to break even.

Common methods for measuring customer churn include the net promoter score (NPS), customer satisfaction (CSAT) surveys, and customer effort score (CES) surveys.

Common ways to reduce customer churn include providing excellent customer service, offering competitive prices, and regularly improving product quality. In addition, companies can use data from churn forecasting models to target at-risk customers with retention campaigns.

Customer churn can be predicted using a variety of methods, including machine learning algorithms, logistic regression, and decision trees.

To predict customer churn, you will need data on past customers who have churned. This data should include features such as the customer's demographics, behavior, and engagement with the product or service.

Common evaluation metrics for customer churn prediction include accuracy, precision, recall, and F1 score.


Einblick is an AI-native data science platform that provides data teams with an agile workflow to swiftly explore data, build predictive models, and deploy data apps. Founded in 2020, Einblick was developed based on six years of research at MIT and Brown University. Einblick is funded by Amplify Partners, Flybridge, Samsung Next, Dell Technologies Capital, and Intel Capital. For more information, please visit and follow us on LinkedIn and Twitter.