The Importance of Churn

Most companies have many analytical outputs. The most common difficulty for companies is identifying the useful, accurate, and clear analytical outputs. Churn is at the very top of a short list of required measurements for every company, yet most companies do not have a clear or actionable churn metric definition.

Churn measurement tells you the rate at which customers have left the business and projects that rate into the future. It is useful as a single, composite rate for the business but even more useful at the segment or product levels. This is where you can see where churn is actually originating. Once you see where churn is originating, you can take action to reduce churn, just as you can add insulation or windows to an area of the house from which heat escapes.

‍

Churn is another word for energy loss. Step 1 is to identify where the energy loss is concentrated.

‍

Tracking Churn

Track churn at the customer level.

At the aggregate level, customers should be grouped into cohorts based on when they first became customers or engaged with the business. The precise grouping that is applied depends upon the rate of customer acquisition, but monthly is often a good place to start. Because the product, market, and world change over time, we can assume that customers in the same cohort were exposed to similar conditions and product features as each other. A customer cohort should be examined as though it is a single customer, having unique attributes.

Without doing any machine learning, churn tracking reveals how many customers we are likely to lose in upcoming periods. This is an extremely powerful output because it tells us how many customers we need to go find to achieve certain growth rates. Knowing the required size and scale of business development activity is key.  

Companies with a small number of active customers struggle to track churn. There are numerous approaches that these companies can take to produce reliable churn rates, often starting with more telemetry on the various interactions that customers have with the product or service.

Use https://yourchurndashboard.ai to track your churn metrics

‍

Predicting Churn

Predict customer churn for surgical precision.

Most companies should not invest in customer churn prediction. It is costly and, more important, often does not achieve the intended results of mitigating churn. Why is this? Even if you know who is likely to churn in an upcoming period, this knowledge is only useful if you can:

1. Do something about it.
2. Take cost effective actions relative the mitigation benefit.

Companies should think carefully about investing in customer churn prediction. Ensure that you have useful customer attributes and activity data that will support an accurate prediction and that you have some mechanisms to keep these customers from leaving.

Only predict customer churn if you have treatments (levers)

‍

Testing Churn Treatments

If predicting churn is useful for your business, it's key to test the treatments.

A|B testing is core to churn treatment. Without a robust A|B testing framework and capability, the churn treatments will be costly and wasteful. Here are some questions to ask before treating churn:

1. If a treatment retains a customer, how many months or cycles does the customer need to stick around for to qualify it as a success?
2. How much time (opportunity cost) is being burned to retain the churning customers?

In many cases, treating churn in this methodical way leads to high value. But, ensure you take a circumspect approach. It is common for marketing and sales teams to spend time treating churn without careful measurement, or by measuring too shallow.

Treat churn methodically

‍

Defining & Tracking Churn

Churn will mean something slightly different for different businesses.

For B2C subscription businesses, such as Netflix, churn is fairly clear: users cancel the subscription.

For B2C transaction businesses, such as an e-commerce or ads-revenue business, churn needs to be proactively defined. We have to define a period of inactivity (no purchase, site visit, etc.) after which a lapsed customer is deemed churned. This can become complex. We advise you to keep it simple to start.

For B2B businesses, whether it be SaaS, consulting, or usage-based, customer engagement is often dictated through long term contracts. The non-renewing, or significant downsizing of these contracts, is churn. This churn should be tracked throughout the contractual engagement period, so it’s key to have numerous churn metrics that proxy the contract renewal, leveraging any & all customer signals available.

‍

Hakuin enables decisions. Data powers actions.

‍