Cohort Analytics
Cohort Analytics
Cohort Analytics
An organization can then strategize and optimize its techniques and tactics to target
specific Cohorts. Cohort Analysis should not be confused with Cohort study in which
Data is broken down into similar groups, whereas in Cohort Analysis, the Analysis is
performed in regards to Big Data and Business Analytics.
Cohort Analysis is also used as a tool to measure user engagement over a period of
time. Data from a given eCommerce platform, web application, or online game is
used for Analysis and rather than looking at all users as a unit, it breaks them down
into related groups.
b. Deep Actionable Cohort Analysis
Techniques like Cohort Analysis help in deriving actionable insights. A social media
platform may have a large undifferentiated dataset of all the users on the platform.
This technique can help differentiate users across multiple user types depending
upon the Time, Engagement, and various other trigger events.
An Example:
Like customers who log on to the platform on a daily basis can be put into one
Cohort, users which use the platform only once a month can be put into another
Cohort, users subscribing for ad-free service can be put into one Cohort. These
different Cohorts can help the organization target the requirements of these users
better and strategize their services and products accordingly. The analysts of the
social media platform would be catered with a visualization of these different groups,
helping them to track the behavior of these users in a targeted manner.
2. Steps to Perform Cohort Analysis
There needs to be some structure to the Analysis to come up with information that is
relevant to improving the product or service being offered by the organization.
Therefore the right set of questions need to be asked and evaluated to ensure that.
Example: For a content platform, the right set of questions could revolve around the
subject of increasing user engagement duration.
The Analysts need to define the metrics on which the Data needs to be evaluated, it
also includes identification of relevant events that can occur and need to be tracked
Example: user engagement, free trial activation, etc. Correlation between different
events and metrics is also used to track certain trends.
These Cohorts could be defined based on the above-defined metrics and a tier
system could be introduced to target these customers better.
Example: Regular users of the platform are provided with gold tier services whereas
one-time users are only given base level benefits on a platform.
Step 4: Performing Cohort Analysis
Finally, perform the Cohort Analysis using Data visualization and identify trends and
patterns across different Cohorts (usually differentiated using different colors).
This method can help target different types of customers and help provide an
optimized experience depending upon the preferences of that Cohort.
The test results are analyzed and strategies based on the insights obtained using
this approach are implemented.
Retention tables show a group of people that visited your site or used your app for
the first time during a certain time frame. They also display the progressive ‘drop-off’
or decline in activity over time for that particular group (a cohort).
For example, the table above depicts a group of users who initiated the“App Launch”
event and their retention over a period of seven days. Dates of Initiation / Initial
launch such as “4/02” and “4/03” are grouped together as Cohorts and their
Retention Rate across time is observed.
● Acquisition Cohorts
● Behavioural Cohorts
a. Acquisition Cohort
This type of Cohort Analysis segments the user base on the basis of the acquisition
date and time of the service or when the user signed up for a product. This tracking
can be performed with different frequencies relevant to the product like daily, weekly,
monthly, etc.
This type of Cohort Analysis segments the user base on the basis of the actions they
undertake while using the application. Special event triggers can be tracked to
understand the behavior of demographically different users.
Example: Such triggers on a Food Delivery app could be the choice of restaurants
and frequency with which customers order food, and for a social media platform, it
could be the pages a user follows or the posts they like.
The majority of organisations use Cohort Analysis to track and improve the Retention
Rate and reduce the Churn Rate of the users of their services or products. But this
technique has wider applications to improve the business.
Example: An online Fashion Store can use this analysis to track the style and size
preferences of different demographics of users and build the inventory accordingly to
maximize sales. Similarly, it can also help identify targeted recommendations and
promotional offers that can be sent to improve customer trust and build loyalty. In a
broader sense, it can identify the correct path for feature adoption and upgrades
across various online platforms.
Cohort Analysis is widely used in the following verticals:
● E-Commerce
● Mobile Apps
● Cloud Software
● Digital Marketing
● Online Gaming
● Website Security
For the purpose of the analysis of the Retention Rate of the users, the metric used is
called Customer Retention Rate which is calculated with the help of the formula
given below.
Other than CRR there are other metrics that can be used to analyze the Data. As
there usually is too much information in such operations, Cohort Analysis can be
used to highlight the important patterns and metrics to reach conclusions and
solutions. These metrics to help differentiate Cohorts are stated below:
● Repeat Rate
● Orders Per Customer
● Time Between Orders
● Average Order Value (AOV)
a. Repeat Rate
This metric helps track the users which interact/purchase/transact with the
company’s business repeatedly compared to the customers that terminate the
relationship with a single interaction with the services/products.
As the name suggests, this metric helps track the number of orders placed per
customer. An increased Orders per customer metric indicates a strong retention rate
for that product/service.
This metric measures the average duration between the orders placed by a
customer. The optimum value of this metric is dependent on the industry, type of
product/service and the business model opted for by the company. This metric can
help send targeted mail regarding promotions to dormant customers and reactivation
/ resubscribing reminders to present customers.
This metric helps in the identification of high-value spenders and loyal customers
and helps in differentiating high-value Cohorts which can be further targeted with
marketing campaigns and promotional upgrades to earn customer trust and build
brand loyalty.
Below is the retention curve of acquisition cohort for an app. It makes easier to see
exactly when most users are leaving your app.
We can see immediately that the biggest drop off occurs after the first day: almost
75% of users stop using the app.
A second significant drop occurs on day 7, falling to 10% before the curve levels off.
Obviously, the onboarding experience needs to be revamped to get users to
experience the app’s core value as fast as possible.
Acquisition cohorts are great for identifying trends and pinpointing where customers
churn, but they don’t provide insights into why they’re leaving.
Each of your users makes countless decisions about how to interact with your app:
using feature Y vs. feature Z, commenting on content vs. passively consuming it,
completing a profile vs. leaving it blank, etc.
Here’s another example to help simplify how behavioral cohorts can uncover
opportunities for improvement. Using a movie ticketing app as an example, we’ll look
at all users who launched the app and made a purchase.
Then we can compare with all users who’ve connected their PayPal accounts to
purchase movie tickets in one click.
This can answer interesting questions, like:
Next, test user behavior by looking at a cohort analysis example that compares
users who purchased tickets and those who abandoned their shopping cart.