Cohort Analysis: Why Every Company Needs It, & Real Life Examples.
Cohort Analysis analyzes cohorts based on a certain user event/behavior across time.
Whats is Cohort Analysis?
Cohort analysis is a type of behavioral analytics, which groups users who share a common characteristic.
Primarily identified by breaking down customers into related groups in order to gain a better understanding of their behaviors.
For example,
- User Acquired by Social Media Ads.
- Paid customers for October.
- Paid Customer by different category of Business(Purchased Mobile, Fashion apparels, Beauty, etc)
How can you use Cohort Analysis to amplify your Business?
Cohort Analysis is the cornerstone of retention analytics and customer retention is crucial for business success.
The misconception about Cohort Analysis.
Just segmenting by the date in which visitors become customers or sign up is useful, but doesn’t give a clear picture of what makes each user different.
Segmenting users by the behaviors taken on your app or website allows you to paint a much clearer picture of how people interact with your product throughout their lifecycle.
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 with different age brackets and further narrowing down basis behavioral events female group who likes discount , a female group who likes new offerings.
By breaking down the group in a meaningful way, businesses can start seeing the trend in the data and take action required.
- Correlation between Discount and LTV
The below table is an LTV comparison of cohorts 25 -34 years Female having the same demographics but different behavior.
As you can see from the above table the LTV for users not acquired using a discount is much higher compare to users acquired using a discount.
2. Correlation between Frequency(Magic Number) and Retention.
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
The magic number can be any behaviors or events that your highly retained users do that differs from the rest of the users.
In the example above as you can see users who have purchased more than 3 times in a month tend to have significantly higher retention in the following month compared to users who have purchased less than 3 products in a month.
3. Correlation between Category and Retention.
One can create a hypothesis that some categories trigger maximum stickiness among users when they are the first purchase.
From studying the chart, one can draw the following conclusions:
- Users buying Fashion Apparels in the first purchase showed higher retention than the rest.
- Users buying Jewelly in the first purchase showed the lowest retention rate.
- 1st month is critical as the churn seems to increase beyond that.
Some possible inferences can be that the marketing expense for Jewelry needs to be decreased. Likewise, the retention strategies for Jewellery purchasers need to be relooked. Retention strategy for users entering the 4th month since their acquisition has to be evaluated.
4. Correlation between Acquisition Channel and Retention.
With acquisition cohorts, you can look at the retention of a specific cohort based on when they started using your website.
This cohort divides user’s retention based on when they were acquired through different channels. Depending on your channel, user acquisition could be tracked daily, weekly, or monthly.
With this analysis it helps you to select best channelsfor your business.
Takeaways
- Determine what question you want to answer.
- Define the metrics that will be able to help you answer the question.
- Define the specific cohorts that are relevant.
- Perform the cohort analysis.
- ’’Test results”
