4 Key Metrics for Predicting Customer Churn
Keeping customer churn low is not a lofty goal for most companies — it's a requirement for continued success. Not only is it less expensive to retain current clients than to acquire new ones, but increasing existing customer business is one of the fastest ways to grow revenue. It's no wonder then, that with the rise of data analytics, marketing and customer success professionals are looking for ways to apply some science to the art of predicting customer churn.
Yet, while the idea of keeping customers may seem basic, analyzing data to determine when and why buyers leave is far from it. First you need to make sure you have the right data and the right amount, not to mention significant domain expertise. If you're not a data scientist, you may need to turn to tools such as business intelligence tools to find correlations and patterns within the data in addition to helping you glean insights that will predict increases in churn.
The trouble, however, is that you have to decide which metrics to consider and correlate — and which to exclude. This is a process fraught with bias because, let's face it, Kevin in Sales is never going to include metrics that show churn is his team's fault. Fortunately, we can help you skip over the part where you focus on the wrong data or expend too much time figuring out which metrics are the best.
Here are four key metrics that have proven to be the strongest indicators of customer retention:
1. Loyalty. Loyalty and retention go hand-in-hand—a loyal customer is one that is more likely to come back to your business. There's a range of key performance indicators (KPIs) that provide insights into this realm. For instance, consider:
- Net Promoter Score (NPS): An NPS is based on your customer's answer to the following question: How likely would you be to recommend our product/company/service? The responses are ranked on a scale from 1 to 10. Customers that score the highest are most likely to purchase more from your company or refer your business to others.
- Loyalty program engagement: From punch cards that provide rewards for a certain number of purchases to a fee for added perks —think Amazon Prime—customers that make use of loyalty programs are usually, well, pretty loyal. Understanding the behaviors of those buyers that are most engaged can help your company determine which customers will stick around — and why.
- Rate of coupon redemption: Monitoring coupon redemption rates is another way to assess loyalty. Whether you're offering mobile discounts via text or more traditional coupons, tracking how often they're redeemed is a valuable way to assess which customers will continue to be fans.
2. Tenure. Metrics that examine tenure are measuring the duration of your customer relationships. They can serve as a canary in the coal mine—if you start losing longstanding customers, you want to figure out why as soon as possible. Data points that speak to tenure include:
- Loyalty program membership length: How long customers have participated in your loyalty program is a key indicator of their commitment to your brand. For instance, Starbucks reports on its “active rewards members," which are people who've used the program for more than 90 days.
- Amount of time as a customer: This is a straightforward metric that measures how long customers stay with your company. Depending on your industry, different variations on this may make more sense. Cloud-based software companies may, for example, measure customer tenure as how long a customer has subscribed, while a retailer may individual track site visits over a defined period of time.
- Customer Lifetime Value (CLV): The CLV refers to how much profit a customer will generate over the course of the relationship with your company. You can compare it to other metrics, such as the cost of customer acquisition, to determine how long it takes to recoup your investment in acquiring a new client.
3. Shopping amount. How much your customers spend is a critical metric, though one that varies widely by industry. For example, shopping amounts between consumer and manufacturing customers could vary by tens of thousands of dollars. Regardless, knowing what to expect from your regular customers is critical. Metrics that highlight shopping amounts include:
- Average transaction size: This reveals how much customers spend with each purchase. It's especially helpful information when paired with customer segments or correlated to seasons—so you can predict when customers may be planning to purchase more.
- Units per customer: This metric shows how many items customers buy during each transaction. Paying attention to units per customer can also help predict seasonal trends as well as provide indicators of customer segments at risk.
4. Shopping regularity. Tracking how often customers make purchases also informs future churn issues. For example, a dip in shopping regularity may signal a pending retention problem. Data points regarding regularity include:
- Purchase frequency: You can determine this by analyzing how often customers make a purchase within a specified period of time, say a year or a month. A high purchase frequency is likely an indicator of high retention—and potentially lower churn.
- Time between purchases: Measure the duration between purchases, which, again will vary by industry. Look for changes in the average duration as a signal that customers may be pulling away from your brand.
Once you determine which metrics within these four retention areas make the most sense for your business, you can use AI tools to automate the data analysis. Keyence's Ki automated analytics platform makes this part easy by not only analyzing customer churn based on your data points, but also by providing recommended actions for reducing it. Your current customers are the lifeblood of your organization. Take advantage of technology to ensure that you can take the right actions to keep them happy—and understand what to do if they're not.
Learn more by downloading our white paper on How to Target Your Best Customers and Keep Them.
LATEST POST : What Is Prescriptive Analytics?