To say that the subscription-based business model is picking up steam is a massive understatement. While not too long ago it seemed to be mainly the purview of telecommunications, software and media companies, today businesses of all sizes are selling groceries, organic produce, meal kits, cosmetics, personal grooming products, health supplements and much more as weekly, monthly or annual subscriptions.
Consumers have enthusiastically embraced the model because of the convenience it offers, while businesses welcome the promise of steady recurring revenues. And just as the pandemic (with its stay-at-home orders) has accelerated the shift to ecommerce, it has also hastened the adoption of retail subscriber models.
If you operate a subscription-based business, you should be concerned about cancelled subscriptions, often referred to as customer churn.
Why should you care about customer churn?
As the owner-operator or manager of a startup or growing businesses, you’re likely focused on customer acquisition. It’s in an entrepreneur’s DNA to strive to grow the business.
You should be just as fixated on customer retention. Why? Because it’s much more costly to acquire new customers than it is to keep current ones — five times more expensive, according to some studies. If you’re continually spending to win new customers, you could be digging a big financial hole to fall into. Reducing by even a small percentage the number of customers who leave you each month can result in a big revenue lift over time.
So, staying on top of customer churn is important for business growth. But with so many other things to take care of, including fulfilling and shipping orders, restocking supplies, managing company finances, updating your digital storefront and more, you may not even know what your churn rate is.
A formula for customer churn
Customer churn = (customers at beginning of month - customers at end of month) ÷ customers at beginning of month
Don’t include new customers in the customers at the end of the month, because you’re trying to calculate the customers you lost.
Why is it important to calculate and report on customer churn? Because it allows you to forecast future revenues more accurately. It’s also a key metric for the health of your business — a leading indicator of the perceived value of your brand and your promised customer experience.
Predicting the future is a science, not a dark art
What if you knew which customers were likely to leave, and you could take action to keep them loyal to your brand? If you have hundreds or thousands of subscribers, this might seem like an impossible dream. But what if this kind of intervention could be automated?
Customer churn prevention is possible — even common in some fields of business — thanks to data science practices like propensity modelling and predictive analytics.
Predictive modelling has been around for a while, and it’s frequently powered by machine learning. It’s often applied to the issue of customer churn.
The key to predictive modelling is data. In the case of churn prediction, it’s specifically customer data: demographic data in your CRM/customer database; usage and transaction data captured from your ecommerce site; customer support records; plus any additional sources of customer insight that you have at your disposal.
Businesses that obsess about customer experience and regularly engage with their customers to capture feedback and preferences are best set up for customer churn prediction. Why? Because they’re likely also sitting on a treasure trove of data that can be used to build a robust prediction model. But even if you’re not in this category, you can still benefit from propensity modelling.
A high-level overview of customer churn prediction
To start, you need to decide which data you’ll use to compare customers who churned (i.e., lost customers) to your current subscriber base to identify those at greatest risk of churning (now or X months from now). You’ll focus on particular “features” such as age, gender, spend, number of support calls, etc. as key indicators of churn.
Next, you’ll need to extract these data sets from the databases they’re in and consolidate them in a usable format for predictive analysis. This is the work of data engineering, and it’s critical to get this right or your predictive models won’t work.
You will have to build a predictive model that will enable you to identify customers at risk based on the historical customer data. This typically involves machine learning — essentially training algorithms to automatically recognize patterns in the data that are indicators of a propensity to churn. While there are cloud-based prediction services to consider, this crucial process often involves bespoke development by data scientists.
Once your data sets are prepared and your predictive model built, trained and tested, you can run the analysis and start putting the results to work in your business.
The value of customer churn prediction
By integrating customer churn prediction into your reporting and analytics, you’ll not only determine which customer segment is most at risk but also when they’re likely to unsubscribe. Then you can make the right move at just the right time to retain them.
You can create an innovative loyalty program that targets a specific group of at-risk customers with a really compelling offer, rather than taking a “spray and pray” approach with a less valuable offer to your entire customer base. You can surprise and delight them with value-added services that help them get more value from their subscription. You can save money on unnecessary discounting and “get one free” offers while achieving better results.
Re-engage at-risk subscribers with your brand before it’s too late, and turn them into long-term customers and advocates for your services.
Go further with customer data analysis
Once you harness the power of customer data in your business, you’ll likely realize that predicting customer churn is just the first step. You can also put your data to use in customer analysis and segmentation analysis that will add power and focus to your marketing, sales and product development efforts.
Most areas of your business can actually be improved with customer data analysis. If you’re early in your journey to becoming a data- and analytics-driven operation, predicting subscriber churn is an excellent place to start. From there you can apply customer insights and analytical capabilities to drive efficiency, productivity and customer centricity across all areas of your business.