
Large companies have millions of customers. Without segmentation, it is very difficult to create marketing strategies or customer communications such as email or newsletters. Customer Segmentation is an important tool for businesses looking to optimize their marketing and product development efforts and better serve their customer.
There are various ways to segment the customer. In this blog, I will focus on transactional customer segmentation, also called as called RFM (recency, frequency, monetary) modeling.
RFM is
R = Recency = When was the last time the customer made a purchase
F = Frequency = How frequently a customer makes a purchase
M = Monetary = What is the amount of purchase made
The process of using transactional segmentation to create a marketing strategy is shown here.

Purchase Transactions
In order to demonstrate transactional segmentation, I will take an example of an online retailer. Sample data of retail transactions are shown here.

The data has information such as the invoice number, the product bought, the quantity bought as well as the price.
RFM Model
The RFM calculations are done at the customer level. The invoice date can be used to calculate the recency, the invoice numbers can be used to calculate the frequency and the total price can be used to calculate the monetary value.
After the RFM calculation, the data will have days since the last purchase, the frequency measured as invoices per month, and monetary value as a total of all purchases in a month.
A sample data on RFM calculation is shown here.

Clustering
The RFM calculation is done for each customer, the time frame could be monthly or yearly or any other time frame required by the business. Using the RFM values, the customers can be segmented using a clustering algorithm.
Shown here is the result of the clustering algorithm. Each dot corresponds to a customer. The color of the dot corresponds to the segment. There are a total of 5 segments and each customer is assigned to a segment.

Segment interpretation
It is good to have a nice-looking clustering visual. However, it is equally important to interpret what the clusters mean.
A visualization that can help in segment interpretation is parallel coordinates, as shown below. You can see a vertical line for recency, frequency, and monetary. All values have been standardized to be between 0 and 100. The rightmost vertical line is for the segments or clusters.

Marketing Strategy
Using the parallel coordinate visualization, we can interpret each segment and develop a marketing strategy. Shown here is a visual for each cluster or segment.

We can now interpret each cluster and develop a strategy.
Cluster 0
- Interpretation— Customers who have not made purchases recently
- Strategy – Make offers to bring them back to purchasing
Cluster 1
- Interpretation – Customers who have a high monitory value
- Strategy – Create a loyalty program so that they can continue spending more
Cluster 2
- Interpretation – Customers who have not made purchases recently
- Strategy – Make offers to bring them back to purchasing
Cluster 3
- Interpretation – Customers who are likely to churn
- Strategy – Retain them with exciting offers
Cluster 4
- Interpretation – Regular customers
- Strategy – Create a loyalty program to keep them purchasing on a regular basis
Conclusion
In this blog, you saw how to go from data to creating marketing strategies for business growth. Customer segmentation using RFM is a very effective strategy that can be implemented using clustering and interpretation visualization.
Datasource citation
The data source used in the blog is available here
https://archive.ics.uci.edu/ml/datasets/online+retail
It can be used for commercial or non-commercial purposes using the following citation
Daqing Chen, Sai Liang Sain, and Kun Guo, Data mining for the online retail industry: A case study of RFM model-based customer segmentation using data mining, Journal of Database Marketing and Customer Strategy Management, Vol. 19, №3, pp. 197â€"208, 2012 (Published online before print: 27 August 2012. doi: 10.1057/dbm.2012.17).
DEMO
You can also watch a demo of RFM modeling on my YouTube channel
Please join Medium with my referral link.
Please subscribe to stay informed whenever I release a new story
Additional Resources
Website
You can visit my website which is a no-code platform to learn Data Science. https://experiencedatascience.com
Youtube channel
Here is a link to my YouTube channel https://www.youtube.com/c/DataScienceDemonstrated