A cohort can be defined as the subset of a large set where the members of this subset have common characteristics. Mostly for business purposes, we put all the customers who joined in the same year/month/date in one cohort and analyze the behaviors of different cohorts over time. This includes analyzing the number of repeat customers, spending patterns, profits, discounts, etc.
Cohort analysis can help us identify high/low performing cohorts and factors responsible for their performance. We can then drill down further to investigate “why” and come up with “what” can be done.
Tableau uses LOD calculations to perform Cohort Analysis very effectively. Let’s take an example to see how it works in Tableau.
For the example demonstrated below, I have used the EU superstore data. The task here is to find out “How many of our customers from previous years contribute to our sales this year?”. To find out an answer to this question we need to do cohort analysis on our customers where the cohort is determined by the year of the first purchase of customers. We have data starting from 2018 to 2021.
To get the cohorts we will use a LOD calculation as shown below:
With this we will get four cohorts based on years:
- 2018 cohort
- 2019 cohort
- 2020 cohort
- 2021 cohort
The contribution by each cohort over years is demonstrated below with the help of a stacked bar chart:
Following stacked bar chart shows the contribution of each cohort in the percentage of total sales for each year:
From the above analysis, it seems that most of the customers are repeat customers and overall superstore is not gaining much of new customers. This gives us an indication that the business might be at risk of experiencing a plateau in a few years.
The above example gives us a peek into the importance and usefulness of cohort analysis. Hope you will now use cohort analysis to get great insights to make an informed and data-driven decision.