Customer Churn Analysis: Understanding Why Customers Leave
- Churn rate = customers lost during the period / customers at the start of the period x 100; customers who joined during the period are not counted.
- A growing customer count can hide rising churn, because new customers replace the ones you lose, so always compare churn rates across periods.
- Churn is a lagging metric: it measures damage that already happened, so pair it with leading indicators such as recency, purchase frequency, NPS and CES.
- Churn analysis needs both quantitative data (who left and when) and qualitative data (support tickets, reviews, surveys) to explain why customers left.
- Active churn (a customer decides to leave) and passive churn (for example a failed payment) have different causes and must be analyzed separately.
Customer churn analysis is the process of finding out why customers stop buying from your store, so you can fix the churn you can avoid. It starts with the churn rate, but it does not stop there: it compares churn across periods and customer groups, then uses feedback and behavior data to explain the numbers.
Sometimes, keeping an eye on the leading indicators of churn is not enough. No matter how proactive you are about churn, you need to perform customer churn analysis regularly. You might be dealing with problems that you did not discover through other types of analysis, and the sources of last year's churn might be different from those driving your churn rate today.
Customer churn analysis helps you stay aware of why customers leave and act effectively. It also helps you improve your product assortment and customer experience, and remove friction altogether. Let's look at its role and its importance to the health of your eCommerce business.
What is customer churn analysis?
Customer churn, or customer attrition, is the metric that shows the percentage of customers who stopped buying from your store during the analyzed period.
Churn can occur at any stage of the relationship: after a single purchase, or after many repeat purchases.
Customer churn analysis helps you understand the underlying reasons behind your store's churn rate. It goes beyond calculating the rate and helps you identify the factors that lead to churn, then use that knowledge to reduce the risk of churning.
Calculating and tracking churn rates should be an ongoing process. Churn analysis reveals anomalies in how your customer database evolves, which you can then investigate through qualitative research. All these steps contribute to better retention in the long run.
In a subscription business, a customer who cancels has clearly churned. In most online stores there is no cancellation: customers simply stop coming back. That is why eCommerce churn needs a definition, usually a period without a purchase that matches your normal buying cycle. Our guide to repeat purchase rate, retention rate and churn rate explains how these three metrics relate.
How to calculate customer churn rate
You calculate the churn rate to know the percentage of customers who stopped buying from your online store in a specific timeframe. The formula is:
Churn Rate = (Number of customers lost during the period / Number of customers at the beginning of the period) x 100
Count only the customers you had at the start. New customers acquired during the period are left out of both numbers. If you include them, strong acquisition can make churn look lower than it is.
How often should you calculate the churn rate? It depends on the types of products you sell and on the customers in your customer base. If you sell something customers buy monthly, it is healthier to track churn monthly. If the average period between transactions is longer than one month, you should at least track the yearly churn rate, and compare shorter periods inside the year.
Worked example. Say your store sells pet supplies. You start the year with 23,000 customers. In the first six months, 575 of them stop buying, and you gain 875 new customers, so you reach 23,300 customers by July. In the second half of the year, 885 of those 23,300 customers stop buying and you gain only 185 new ones, so you end the year with 22,600.
| Period | Customers at start | Customers lost | New customers | Customers at end | Churn rate |
|---|---|---|---|---|---|
| January to June | 23,000 | 575 | 875 | 23,300 | 575 / 23,000 x 100 = 2.5% |
| July to December | 23,300 | 885 | 185 | 22,600 | 885 / 23,300 x 100 = 3.8% |
If you stopped measuring in the middle of the year, you would see a growing customer base and conclude that your marketing is paying off. You would not notice that churn rose from 2.5% to 3.8% in the second half. Note also that the drop in total customers over the year, from 23,000 to 22,600, is not the churn rate: it mixes lost customers with new ones.
By conducting a churn analysis, you can see beyond the tip of the iceberg and identify the causes of the higher churn rate. For example, one cause could be the active churn of dissatisfied customers who did not receive proper treatment from your customer support team. Another could be passive churn caused by a technical payment problem that your marketing and development teams did not notice in time.
Are you struggling with many at-risk customers? Learn how you can reactivate customers who are about to leave you.
Why customer churn analysis matters
The ugly truth about not meeting what your customers need and expect from your brand is that they will not necessarily complain before leaving. Most customers leave silently after one or more purchases, so measuring churn means measuring damage that already happened.
Many companies with new customer acquisition as their top priority do not notice churn until they deal with reduced revenue and decreasing profit. When a company reaches this point, it is already in big trouble.
There are multiple leading indicators of churn that you should monitor, such as purchase frequency, customer satisfaction scores, engagement metrics, and reviews. Churn analysis lets you evaluate the current situation and build a plan for churn prevention that keeps your business away from worst-case scenarios.
By calculating the churn rate for different timeframes, you can identify the problematic periods. Then you can go in depth with a qualitative analysis that helps you understand the "why". Customer churn analysis works best when you ask questions like these:
- Why did we have a great BFCM campaign, but churn rose sharply the following month?
- Why did we lose 20% of our loyal customers in the last 12 months?
- Why did we receive 350 complaints in the previous month alone?
- Why did we lose 15% of our subscription customers in January?
- Why did customer support receive a low score although we doubled the team?
- Why did our customer loyalty scores drop in a specific month?
By performing churn analysis, you keep your company away from dangerous assumptions. You look at your churn data instead of constantly reprioritizing your actions around random tactics to reduce churn that you found online. Analyzing customer churn helps you find the root cause of your problems and make informed decisions.
You can also test your hypothesis, see how customers engage, and measure the impact of your test before you implement a tactic on a larger scale.
High churn rates show that improving your customer retention strategy should be your top priority. Increasing retention raises Customer Lifetime Value, and that is how you produce the budget to secure your business's future growth and fund your acquisition campaigns. For proven tactics, read how to prevent and reduce customer churn.
How to run a customer churn analysis
-
Define churn for your storeDecide how long a customer can go without a purchase before you count them as churned. Base it on your normal buying cycle, not on a generic number.
-
Calculate churn for several periodsUse the formula above for months, quarters or half-years, and compare them. A rising trend is the signal to investigate, even when your total customer count grows.
-
Break churn down by segment and cohortCompare churn across acquisition cohorts, channels, products and RFM segments. Find out which customers leave, and when in the relationship they leave.
-
Separate active and passive churnCheck payment failures, delivery issues and technical errors separately from customers who chose to leave. The fixes are different.
-
Find the reasons with qualitative dataRead support tickets, negative reviews and survey answers from the segments with the highest churn. Look for the problem that the numbers and the feedback both point to.
-
Test a fix, then monitorTest your hypothesis on a smaller group, measure the effect on churn, and roll it out if it works. Keep tracking the leading indicators so you catch the next problem early.
Challenges of customer churn analysis
In their effort to understand and prevent churn, many companies use a predictive churn model. These models use existing customer and eCommerce data to predict how churn will evolve and to show what your teams have to do to prevent it. One study of aftermarket operations, for example, used Recency, Frequency and Monetary value to set churn thresholds and predict future churners (Briker et al., SMU Data Science Review, 2019). To learn how to score churn risk, read our guide to churn propensity. As helpful as it is, churn analysis can also be challenging, for the reasons below.
You need someone who can see beyond the numbers
You could have many loyal, engaged customers who buy from you frequently and regularly, and some of them suddenly stop buying.
Sometimes, the truth behind customer attrition lies in negative feedback or customer support tickets. Not everything your customers say and do can be measured.
That is why churn analysis needs accurate, up-to-date data and someone who can see beyond the numbers. Raw numbers will not help you understand customer behavior, so you need to mix quantitative and qualitative data. While this can take a lot of time and resources, it will definitely cost you less than ignoring a rising churn rate.
You need a unified view of customer data
If you do not use a customer data platform that aggregates your data automatically, it is tough to gather the data manually and get accurate results.
Having all the data in one place is important for any analysis you perform for your store, including cohort analysis and customer segmentation analysis. A tool that unifies customer data is fundamental for optimizing customer experiences, which play a critical role in the success of your customer retention strategy. You cannot identify problems and take the proper measures to prevent churn without complete, updated data. This applies to all types of churn you measure.
Nexus by Omniconvert uses your Shopify purchase history as its primary data source and adds GA4 for acquisition source attribution. It tracks Customer Lifetime Value by acquisition channel and by monthly cohort. If you want to see it with your own data, book a call.
You need continuous monitoring of customer behavior
For effective churn analysis, you need to cover all the leading churn indicators that apply to your business. Some are historical metrics, like purchase recency, frequency, and monetary value. Others are predictive metrics, such as Net Promoter Score or Customer Effort Score.
| Indicator | Type | What it shows | How to read it |
|---|---|---|---|
| Recency | Historical | Time since the customer's last purchase | Recency growing past the customer's normal buying cycle is an early churn warning |
| Purchase frequency | Historical | How often the customer buys | A drop in frequency among loyal customers needs investigation before they stop buying |
| Monetary value | Historical | How much the customer spends | Shows how much revenue is at risk if the customer leaves |
| Net Promoter Score (NPS) | Predictive | How likely customers are to recommend you | Falling scores or more detractors can signal churn before purchase data does |
| Customer Effort Score (CES) | Predictive | How easy it was to get something done | High effort at a touchpoint, such as returns or support, points to friction that drives customers away |
Behavior segmentation, such as RFM segmentation, helps a lot in churn analysis, because it identifies at-risk customers and lets you create the right prevention campaigns for them. In Nexus, for example, the About-to-Dump segment holds previously active customers whose recency is dropping: they are on their way out, but they have not left yet. Nexus syncs segments to Klaviyo, Meta Ads and Google Ads, so you can reach these customers with win-back campaigns.
It is even better to combine behavior segmentation with customer satisfaction metrics and keep a proactive approach. Otherwise, you only react when your customers' trust is already decreasing. For more metrics to track, see our list of customer retention metrics.
It might take some time to identify all the metrics that help you evaluate churn and to design an effective churn analysis model. Still, your company needs to perform this analysis regularly to stay aware of the "why" behind your churn rate.
To learn how to keep a proactive approach and keep churn as low as possible, join our world-renowned instructors in CVO Academy.
Frequently Asked Questions
Customer churn analysis is the process of finding out why customers stop buying from your store. It goes beyond calculating the churn rate: it compares churn across periods and customer groups, then uses customer feedback and behavior data to find the causes. The goal is to fix the churn you can avoid.
Divide the number of customers you lost during the period by the number of customers you had at the beginning of that period, then multiply by 100. Do not count customers who joined during the period. For example, if you started a quarter with 10,000 customers and 400 of them stopped buying, your churn rate for that quarter is 4%.
It depends on how often your customers buy. If customers buy something from you every month, track churn monthly. If the average time between purchases is longer than a month, track churn at least once a year, and compare shorter periods inside that year to see trends.
Online stores rarely get a cancellation signal, so you need a rule. Most stores treat a customer as churned when they have not bought within a set time, based on their normal purchase cycle. A store that sells monthly consumables can use a much shorter window than a store that sells furniture.
Combine numbers with feedback. Churn rates by period and segment show where and when customers left. Support tickets, reviews, and satisfaction surveys such as NPS and CES show why. When both point to the same problem, such as a payment failure or poor support, you have found a root cause you can test and fix.
Monitor the leading indicators of churn: purchase frequency and recency, customer satisfaction scores, engagement metrics, and reviews. A customer who used to buy often but whose recency is dropping is at risk. RFM segmentation groups these customers automatically so you can act before they leave.
Active churn happens when a customer decides to leave, for example after a bad support experience. Passive churn happens without a decision, for example when a payment fails for a technical reason. They have different causes and different fixes, so analyze them separately.
Nexus by Omniconvert uses your Shopify purchase history to build RFM segments, including an About-to-Dump segment of previously active customers whose recency is dropping. It also tracks Customer Lifetime Value by monthly cohort, and syncs segments to Klaviyo, Meta Ads and Google Ads for win-back campaigns.
Churn rate is a lagging metric, but analyzing it can reveal problems your teams did not notice while they evaluated individual campaigns or strategies. Customer churn analysis finds the reasons behind relationships that end too soon, after only one purchase, or too suddenly, ending a long and valuable relationship with a customer who used to be loyal. Calculate churn correctly, compare it across periods and segments, combine the numbers with what your customers tell you, and test your fixes. Do it regularly, because the reasons customers leave change over time.
See which customers are about to leave
Nexus by Omniconvert builds RFM segments from your Shopify purchase history, including the About-to-Dump segment of customers whose buying is declining, and syncs them to Klaviyo, Meta Ads and Google Ads so you can act before they are gone.