A cohort is a group of people who share a common characteristic over a certain period of time. For example, let’s look at a group of students. Cohort analysis allows us to identify relationships between the characteristics of a population and that population’s behavior. Cohort analysis helps you dive deep into your data by analyzing audience behaviors and trends. It’s a powerful and sophisticated reporting technique that helps you boost your understanding of different audience segments.

HOW DOES IT WORK?

A cohort is a group of people with shared characteristics or common criteria. It could be the first article or video they watched on your site or the country/region they are from. By having this element in common, that cohort becomes an audience segment you can analyze and compare with other cohorts to gain deeper insights into content performance across multiple customer segments. Things like viewing patterns, conversion trends, and content popularity, for example.

HOW TO READ A COHORT ANALYSIS CHART?

In this example, we are tracking audience retention from events. The grey column ‘0’ depicts the number of attendees who attended an event each month – it could be a single specific event or all of the events that month. The subsequent columns show the number of attendees who then returned for another event and which month they did so. Reading along the top row of data, we can see that of the 2,560 people who attended events in January, 1,059 went on to attend another event one month later, 794 went on to attend another event two months later, 596 went on to attend another event three months later, and so on. From this, we can immediately see not only which are our best months for acquiring new event audiences (December and January), but also which months/ events have performed best when it comes to audience retention. However, things get more interesting when we start to read down the columns. In the chart above, we can see there is a big drop in numbers for most months between columns 3 and 4, noticeable by the distinct change in shading between those columns. What this is telling us is that three months after the first event is a critical time for audience retention. Armed with this knowledge, a business can now start to address why this is, and, for example, develop a loyalty campaign around the 3- month mark to draw audiences back in You can also read the data diagonally. If we look at the July 2020 row, there is a slight jump and this trend continues diagonally up and right across the chart. The June row sees a jump in additional sales in month 2, May sees a jump in month 3, April in month 4, and so on. Clearly, something happened in June 2020 that drew audiences back in. Perhaps this was a very popular event with a wide general appeal or, it might have been due to some external factor that the business will need to determine so they can adjust their strategy to capitalize on similar opportunities.