Cohort Analysis: Definition & real-life examples
Table of Contents
- What is a Cohort Analysis?
- The misconception of Cohort Analysis
- Why is Cohort Analysis important for businesses
- Understanding Cohort Analysis
- Real-life Examples of how cohort analysis is used with marketing strategies
What is a Cohort Analysis?
Cohort analysis was initially introduced in Social Science as a data analysis technique for analyzing the behavior of a group of similar individuals (Cohort). In the business world, it’s used continuously by E-Commerce sites and mobile App to break users into smaller groups to analyze and understand a group’s behavior.
Google define cohort analysis as:
“Examine the behavior and performance of groups of users related by common attributes.”
The misconception of Cohort Analysis
If you do a Google Image search of Cohort Analysis, you probably have seen something like this. But this is not how Cohort Analysis usually looks like. As cohort analysis can be represented in many other ways, this is the most frequently used format by analytic tool companies. As the definition suggests, Cohort Analysis is more of an analytic method than a graph design.
Then why do most images results show something similar to above when “Cohort Analysis” is searched? This misconception might be due to the popularization of the Cohort Analysis Report in Google Analysis, which is by most websites on the internet.
That said, the table above is nevertheless a Cohort Analysis that analyzes cohorts based on a certain user event/behavior across time.
Why is Cohort Analysis important for businesses
Instead of looking at the data as a whole, cohort analysis allows businesses to analyze the behavior of a group of people by first breaking them into smaller groups. A simple example of Cohort Analysis will be breaking users into male and female, analyzing how males react when they look at a travel website when compared to the female counterpart. Cohort analysis can also further narrow down the gender group into, a female group who likes adventure, a female group who likes luxury travel and a female group who likes traveling with family,
By breaking down the group in a meaningful way, businesses can start seeing the trend in the data and take action if accordingly if required.
Understanding Cohort Analysis
One of the ways of doing cohort analysis is pairing it with different analyses such as retention analysis or magic numbers analysis. Due to the rise of the Mobile App industry, retention analysis is one of the most common ways to use cohort analysis to understand how a certain cohort group behaves within the App environment.
Cohort Analysis with the Retention Table
The retention table or retention analysis can be used to analyze when a user executed a specified action, and the retention of the said user overtime.
For example, the graph above showed a group of users who initiated the“App Launch” event and their retention over a period of seven days. Dates of initiate launch such as “4/02” and “4/03” are grouped together as cohorts and their retention rate across time is observed. For the table above, you can notice that:
- Cohorts from 4/06 and 4/07 have a relatively higher retention rate when compared to other cohorts. Further analysis can be done to understand why this difference exists
- The Cohort from 4/15 has a relatively low level of App Launch after one day. But after further observation, it seems to be more of a systematic decrease in retention rate on 4/15. This might indicate a data, server, or App error on that day.
To up the cohort analysis to another level, you can also add what we call the “Magic Number Analysis” into the mix.
Magic Number Analysis
So what is the magic number analysis?
“The Magic Number Analysis originated from Twitter when they discovered that new users who followed at least five users when they registered ended up having a significantly higher retention rate than those who followed less than five users. Due to this, Twitter created a user flow that prompt newly registered users to follow at least five users before they can start using their services. To this day, the same flow is still used in Twitter’s user registration process.”
In other words, Magic Number Analysis, as the name suggests, is to discover the events that occur that lead to an increase in retention rate. This event could be anything from “Users who viewed certain content for more than 7 times” or “Users who launched our App three times a day.”
To learn how you can conduct the magic number analysis, you can read the article here
Real-life examples of how cohort analysis is used with marketing strategies
Example 1: Media for highschool girls - HaruHaru
One of our clients, HaruHaru is a media company targetted at high school girls in Japan. When they first launched their App, they have difficulty retaining their users. After conducting the Magic number analysis, they discovered that users who are logged for three days consecutively had shown higher retention rate when compared to the Cohort that did not login for three days in a row.
To tackle this issue, HaruHaru devices a three consecutive login campaign, which resulted in a 10% increase in retention rate.
Read more about HaruHaru’s case study here.
Example 2: Vehicle listing site - Goo-net
Another client of ours, Goo-net, one of the largest used car listing sites in Japan has a very similar issue with HaruHaru. After conducting the retention analysis, they found a high drop-out rate for new users who downloaded our App but did not conduct any search. To tackle this issue, they came out with a marketing campaign that is aimed at educating users on how to use the search function.
As a result of this marketing effort, they saw a 30% increase in page views.
Read more about Goo-net case study here.
In Conclusion, Cohort Analysis is a compelling analysis to let businesses to take break users down into a smaller group to make data more understandable and actionable. When paired with marketing, it can bring measurable results instantly.