RFM analysis using the Tile Tool w/ "Equal Records" method

[cperazza]

Hi,

Probably a newbie question, but let's see if I can borrow your brains for a minute to address my issue with the Tile Tool.

I'm currently working on a workflow to calculate and attribute a score to customers based on the RFM analysis, where you collect data from all the transactions made by a customer over a period of time, and score them based on the Recency, Frequency & the Monetary value for the business.

In the RFM analysis, I have to split my customer base into deciles (we chose deciles to represent our strategy best), based on those variables above (the R, the F & the M). I chose the Tile Tool, with the "Equal Records" method to bin my customer into deciles 1 to 10.

The issue that I'm facing with the Tile Tool is that regardless of the way I sorted the Recency, the Frequency and the Monetary columns of my dataset, the tool seems to "automatically" sort all those columns in a descending way. This is messing the calculated scores because all the scores have to follow the same criteria: the 1st decile is the lowest score band, while the 10th decile is the highest score band.

When it comes to the Frequency (total orders) & the Monetary (total revenue) values, the higher the value of these columns for a customer, the higher the decile this customer should be put in (people who buy more & more frequently are more valuable for the business. Therefore they have a higher score). But when it comes to the Recency (days since last order) the calculation has to be made oppositely: the lower the number of days since last order, the higher the decile this customer should be put in (a customer who has purchased last week requires different actions compared to a customer who has purchased 5 months ago).

To address that, I made sure that I sorted the RFM columns of my dataset (accordingly to the criteria above) before moving to the step of the Tile Tool. Still, the Tile Tool always set the first decile to the highest values of the columns (RFM) regardless of my pre-sorting step on the workflow.

At the moment, I have to use a Formula Tool to invert the decile attributed by the Tile Tool for the Recency column, but I'm probably missing something in regards to the way that I'm setting the Tile Tool.

The question is:

Is there a way of avoiding this extra step with the Formula Tool, or to force the Tile Tool to read my dataset in the way I sorted?

Kind regards,

[atcodedog05 · Accepted solution]

Hi @cperazza 

Its Tile tools default built functionality it will sort the data before grouping. There is nothing that can be done about it. There no config to change it. Except for a hack like your

Hope this helps 🙂 

[cperazza]

Hi @atcodedog05

Thanks for your answer. I guess I'll have to leave the workflow as it is.

Cheers!