Purchase Frequency Stratification

[MLemanski]

Hello,

I am trying to solve the following problem in PowerBI.  I am analyzing customer purchase frequency, so I created a dynamic measure that calculates how many unique months a customer purchases an item in over a period of time.  I then divide that by the months that are in that period to come up with a Customer Purchase Frequency.  

[Customer Purchase Frequency] = [Unique Months Customer Purchased] / [Months in Selection]

This results in a percentage that I can then put into buckets, no matter the length of the period selected.  For example:

Purchase Frequency	Min	Max
0 - 3 Months	75.0%	100.0%
3 - 6 Months	50.0%	75.0%
6 - 12 Months	8.3%	50.0%
12+ Months	0.0%	8.3%

What I would like to do now is summarize my sales, unique customer count, etc. by my "Purchase Frequency" buckets as defined above.  The issue I am having is that since my "Customer Purchase Frequency" is created from a measure, I can't use it or my buckets based off it as an axis on any charts or tables.

I believe the solution has to do with using a disconnected table, formatted like the table above, but I am not sure how to get my sales, customer counts, and other calculations to work with it.  Any direction would be much appreciated.  Also, if there is a simpler solution to this problem, please let me know.

Thanks!

[v-juanli-msft]

Hi @MLemanski

If you want the "Customer Purchase Frequency" to be used in axis, you could create a calculated column with theis formula.

If you want to get a table using the column from your original table, you could consider the dax function Summarize.

Best Regards

Maggie

[MLemanski]

Hi Maggie,

My issue is that the Customer Purchase Frequency is a dynamic calculation.  A customer's purchase frequency may be "0-3 Months" for one category of product, but "6-12 Months" for another category.  When I try using a calculated column, it just gives me an overall frequency, regardless of how I want to filter it.

I have my data in the standard format, with "Sales Data" table connected to a "Customer" table, "Item" table, and "Calendar".

In PowerPivot, my work around was creating a pivot table that listed each customer and calculated their purchase frequency and sales based on which item category slicers I selected.  I then referenced that table using Excel's native CountIfs or SumIfs functions to summarize them into the buckets that I wanted based off the logic in my original post.

I hope this adds some clarity for what I am looking for!