What is a cohort?
A cohort usually refers to a group of customers acquired at the same point in time (e.g., the May cohort refers to the users or customers who were acquired in May).
In the context of cohort analysis, one can compare the behavior of one cohort with other cohorts. This often involves looking at customer retention and analyzing how the number of customers in a cohort evolves over time.
The comparison helps to understand whether, for example, younger cohorts have better retention, i.e., remain active longer (e.g., because the product has improved over time), the startup has found new marketing channels through which it acquires "better" (i.e., stickier) customers, or similar.
What is customer retention?
Customer retention is an essential performance metric for startups. It tells you how sticky your product or service is and accordingly provides insights on whether your user or customer acquisition is sustainable in light of the expected revenues or contribution margins. Of course, this is especially true for companies whose model is designed or reliant on repeat business.
Customer retention example
A direct-to-consumer company selling mattresses is likely not designed for repeat business because the average customer only buys a new mattress every few years. So ideally, profit is made with each customer on their first order. It's different with Uber, for example - here, the customer usually doesn't become profitable with the first ride but only over time.
What do good retention and cohort look like?
What good retention or cohort should look like is—as described above—highly dependent on the business model. In order to be able to evaluate and classify the retention of a startup, a look at the Customer Lifetime Value in relation to the CAC (Customer Acquisition Cost) or a consideration of the Payback Duration (how long does it take per customer to earn the marketing costs?) is helpful.
You will find here an example of a cohort analysis with explanations.