Hello! In this post I am going to explain how to create a Cohort Analysis.
A cohort is a group of customers or subjects that have a common characteristic. Cohorts can be determined by the date at which they became a customer, by their age, demographic, or any other feature that could be used to group a set of people. It is then assumed that something about this cohort drives certain behavior over time.
You can adapt this explanation to your particular use, but for this case I will create a yearly analysis of customers using Sample-Superstore
Imagine this situation: you have a dataset with information about Customer Names, Order IDs, Sales etc. But instead of analyzing yearly sales looking at all your customers as one single group, you want to break them down into cohorts to have a better understanding of their behaviour.
Therefore, you want to identify the number of new customers for each year in order to analyze their contribution to the yearly sales of the company and to discover insights such as:
- In year 20xx the majority of the revenue was made by customers acquired in the first year of the business (your loyal customers ☺ )
- Or if on the contrary, each year you have a bunch of new customers, but they don’t buy again the next year, so your retention rate would be low…
The key word here is acquisition of the customer i.e the first order made by a customer:
{ FIXED [Customer Name]: MIN([Order Date])}
By dragging Order Date to Columns, Sales to rows you would have the yearly sum of sales, if you increase the level of detail by adding this new field: Customer Acquisition Date, to the colour mark you could now identify that, for example, in 2018 almost ¾ of the sales were from customers acquired in 2017.
Lastly, if you prefer this analysis as a percentage, right click in Sum(Sales) to convert it in a table calc.
If you want to enrich this analysis you could also add a dimension like Category or Region to the filters.
I hope you find this useful!! :)