Entering StartUpLand: An Essential Guide to Finding the Right Job

by Jeffrey Bussgang

Another important tool in figuring out unit economics is conducting cohort analysis—organizing customer behavior data into related groups and analyzing each group, or cohort, to determine trends over time. Examples of common cohort analysis include tracking customers who began using a service at a given point in time and seeing how they behave month over month, or how the customers who were acquired via a particular source perform month over month. Suppose, for example, it costs you ten cents in advertising and marketing costs on Facebook to drive an installation of the popular augmented reality gaming application, Pokemon Go. When you do your cohort analysis on the Facebook-acquired users, you observe that they never actually use the product—they just install it and abandon it—and thus are worthless to you. Then, suppose you launch another customer acquisition initiative on Snapchat where each user costs you twenty cents to acquire, but when you do your analysis of this cohort of users and how they behave over time, you observe that these users play Pokemon Go incessantly—twenty times a day, every day, month after month after month. Thus, your cohort analysis shows that the Snapchat-driven acquisitions are far more valuable than the Facebook-driven acquisitions.