RFM Using Tableau (Study Case: Ecommerce Dataset)
Customer segmentation is the practice of dividing a customer base into groups of individuals that are similar in specific ways. One of the techniques of customer segmentation is RFM (recency, frequency, monetary). RFM segmentation allows marketers to target specific clusters of customers with communications that are much more relevant to their behavior.
Purpose
Create effective allocation of marketing resources and the maximization of cross and up-selling opportunities.
How is the customer recency, frequency, and monetary?
About RFM
Making RFM requires 3 aspects, namely recency, frequency, and monetary.
· Recency is the maximum time an order occurs overall — time_limit each customer + 1
· Frequency is the number of times a customer places an order
· Monetary is how much the total expenses per customer
Handling Outliers using Python
Handling outliers using python, then outliers on recency and monetary are removed. This is because the outliers on the two attributes are few. Meanwhile, if we remove the outliers frequency, there will be only frequency 1. So I don’t remove outliers in frequency.
· The quantiles will be used as a divider for the RFM segment, except for the frequency.
· RFM is divided into 4 segments where The segment with the lowest value is 1 and the highest is 4.
· For monetary, segment 1 is for customers that spend very little money, while segment 4 is for customers that spend so much money.
· The recency will be reversed. So segment 4 is for customers who recently come to orders, meanwhile segment 1 is for customers who haven’t orders for long periods.
· For frequency, segment 1 is for customers who shopped 1 time, segment 2 is for customers who shopped 2 times, segment 3 is customers who shopped 3 times, and customers who shopped more than or equal to 4 are segment 4.
Build RFM Group
RFM Segments build from quantile that can be seen using function describe in Pandas.
After forming 4 segments, these segments are grouped into RFM Group. Each rfm group has a different name and marketing strategy.
The visualization of RFM Group Overall
We can see that most of the customers are Regular Customers. We will deep down analysis on each RFM Group (except Lost Cheap Customers and Best Customers since they are assigned into specific RFM Segment).
Regular Customers
Traits of most Regular Customers:
· Many customers (33,81%) spend a lot of money (monetary = 3) but have a very small frequency (frequency = 1).
· Most of them (49,31%) have a distant last shopping time (9 months — 1 year ago) (recency = 2).
· Although many of the regular customers are in the monetary 3 category, their average spending is still below the overall average and 2 from the bottom.
· Suggestion:
Make limited-time offers, send personalised emails, offer personalized recommendations
Lost Customers
· Most customers (48,31%) spend quite a bit of money (monetary = 2) and very little shopping frequency (frequency = 1).
· Some customers (96,42%) have not shopped for more than 1 year (recency = 1) and only one time (frequency = 1).
· The average spending of “Lost Customers” is below average.
· Suggestion:
Send personalised emails, it is better to ignore it because attracting customers who have not shopped for more than 1 year will require more effort and cost a lot of money.
Recent Customers
· Many customers (34,27%) spend a lot of money (monetary = 3) but have a very small frequency (frequency = 1).
· Shopping these days (recency = 4) but most of them (96,97%) still have less frequency (frequency = 1).
· Although many of the recent customers are in the monetary 3 category, their average spending is still below the overall average and 3 from the bottom.
· Suggestion:
Gift them discounts/promo.
Loyal Customers
· Many customers (73,58%) spend a quite a lot of money (monetary = 4).
· Most of them (56,60%) have a distant last shopping time (9 months — 1 year ago) (recency = 2).
· Their monetary is very good and above average.
· Suggestion:
Offer personalized recommendations, upselling, look back at marketing strategy on 9 months — 1 year ago.
Customers That Need Attention
I crossed between frequency and monetary. 31.33% of customers spend quite a lot of money but only make one purchase. If examined further, most of them have stopped shopping since 5–9 months ago.
These customer segments are potential big spenders and need attention big spenders. The difference is the recency segment of Need Attention Big Spender is1–3, meanwhile, the recency segment Potential Big Spender is 4.
Why Potential Big Spender & Need Attention Big Spender?
- There are quite several them, it will be very profitable if they can level up to become best customers.
• Can shop with a large nominal
· Suggestion:
Conduct in-depth analysis of events 5–9 months ago.
Potential big spender: offer personalized recommendations, encourage them to buy products more frequently so they can level up their member status with so much benefit.
Need attention big spender: reach them via email/notifications, make subject lines of emails very personalized, revive their interest by a specific discount on a specific product.
Summary
This is the summary from the solution above.
Thank you for reading!
