Most SaaS founders pray to the [[Churn Rate]] [[product metric]], which shows how many customers dropped off during the current month. It is calculated as the ratio of the number of customers who left during the month to the number of customers at the beginning of the month, expressed as a percentage. However, this metric is directly related to the efficiency of the business only if we assume that the probability of a customer leaving does not depend on the time for which he has been subscribed to the service. In reality, this is not the case - old users leave less often, and new ones - more often. That is, the probability of leaving decreases as the length of the previous subscription increases. With such a decreasing probability curve of churn, calculating the churn rate no longer signals the health or illness of the business. The churn rate may fall even as the business itself is plummeting, or it may remain unchanged even as the business is growing well. Let's take the simplest situation - you have launched a mass advertising campaign. A lot of new users have come, but many of them dropped off in the first months. The churn rate has increased significantly, but some users stayed and continue to live with us. The business is growing, although the churn rate is alarming. Now imagine that we are conducting such campaigns constantly. The churn rate will fluctuate, but in reality, everything is fine. > [!NOTE] > It turns out that the growth of the business is not affected by the churn rate, but by the shape of the same lifetime distribution curve of users in the service. The simple characteristics of such a distribution curve are the mean lifetime (the arithmetic mean of all lifetimes) and the median (the lifetime of a typical user). At first glance, it may seem that the critical parameter is the median. However, the critical parameter is the arithmetic mean. Let's take two distribution graphs with the same median but different arithmetic means. The graph with the higher arithmetic means implies that it has more active churns of recently subscribed users and less active – of older ones. This leads to an increase in MRR several times compared to another form of distribution. In short, the main money for subscription services comes from a small number of users who live longer than usual - who increase the average life span. If we bring in a bunch of new users, we get their crazy churns in the first months of life - forget about it. > [!NOTE] > The growth of the business is not influenced by the majority of those who leave - but by the minority who remain.