Focus on the Highest Points of Leverage
Once you have established an approach to retention and a baseline performance metric, we recommend taking a thoughtful approach to segmentation in order to gain deeper understanding of retention drivers. Often, teams will begin by segmenting based on readily available attributes, irrespective of usefulness. We recommend thinking more deeply about what drives behavior and what insights can realistically be addressed in your business processes. For example, considering behaviors such as frequency and recency of interaction can help identify opportunities for personalization, which is especially useful for ecommerce or retail businesses. Demographic segmentation is often important for B2C businesses looking to understand the persona of their customer base; this approach may be less useful for B2B companies and can create concerns around privacy or bias for employee segmentation. For mobile products, it may be helpful to segment your users based on platform or operating system; this data may be less relevant for other businesses.
In addition to identifying potential segments with controllable levers, confounding factors should also be carefully considered in order to control potential variables that can impact the analysis. These variables are typically not part of the segmentation criteria but can affect the interpretation of the results. For consumer businesses, seasonality, such as holidays or a major cultural event, may influence customer behaviors across different demographics. Correlated variables, such as age and income, can also confound segmentation results. For example, a business may find older customers or customers using iOS instead of Android devices, to have higher retention rate. However, this may actually be due to their higher income levels, which is the true driver for retention.