Do you know how your Adwords-acquired customer segments grow their LTV compared to those customers acquired from organic search? Have you ever thought of performing a cohort analysis on different customer segments side-by-side in the same report? If so, a qualitative cohort analysis will help you answer those questions.
In this article, we’ll dive into what a qualitative cohort even is, why you might be interested in building this analysis, and how you can create it in Magento BI.
Cohort analysis in general can be broadly defined as the analysis of user groups that share similar characteristics over their life cycles. It allows you to identify behavioral trends across different user groups.
For a more in-depth primer on cohort analysis, take a look here - we wrote the site on it!
Most cohort analyses in Magento BI group users together by a common date: i.e, the set of all customers who made their first purchase in a given month. A qualitative cohort is a little different: it’s a user group that is defined by a characteristic that isn’t time-based. Some examples might be:
The Cohort Analysis Builder is optimized for grouping cohorts using a time-based characteristic. This is great for analyses focusing on a specific segment of user, i.e. all users who were acquired via a paid search campaign. In the Cohort Analysis Builder, you can (1) focus in on that specific user group, and (2) cohort on a date (like their first order date).
However, if you want to analyze the cohort behavior of multiple user segments in the same cohort report (paid search vs. organic search vs direct traffic, perhaps?), this more advanced analysis can be constructed in the Report Builder.
Creating a qualitative cohort report in the Report Builder involves our analyst team creating some advanced calculated columns on the necessary tables.
To build these, please submit a support ticket (and reference this article!). Here’s what we’ll need to know:
Once our analyst team responds to the above, you will have a couple of new advanced calculated columns to build out your report! Then you’ll be able to follow the below directions to do this.
First, you’ll want to add the metric you’re interested in cohorting, once for each cohort you are analyzing. In this example, we want to see cumulative Revenue made in the months after a customer’s first order, segmented by the User’s referral source. This means that, for each segment, we will add one Revenue metric and filter for the specific segment:
Second, you should make two changes to the time options of the report:
In our example, we’ll be looking at an all time view of Revenue. After this, you should end up with a series of dots:
Third, you will make an adjustment to actually set up the cohorts. Based on the cohort date and time interval you specified to our analyst team, you’ll have a dimension in your account that will perform the cohort dating. In this example, that custom dimension is called Months between this order and customer’s first order date. Using this dimension, you should:
Now, you’ll be able to see one line for each cohort that you specified. Check out our example now -- we see the Revenue contributed by users of each referral source, grouped by the number of months between their first order and any subsequent order. We also added a Cumulative perspective to see the cohorts' aggregate growth - take a look at the results table for more granularity.
What does this tell us? Here, the specific referral source Paid search is very valuable in the first month of a customer’s purchasing lifetime, but fails to retain its customer base with repeat revenue. While Direct Traffic starts off at a lower amount, revenue in subsequent months actually accumulates at a similar pace.
No matter how you dice it, cohort analysis is a powerful tool in your analysis toolbox. This type of analysis can yield some very interesting insights about your business that traditional time-based cohorts may not, enabling you to make better data-driven decisions.