This algorithm enriches the customer profile with data and scores about their purchase Recency, Frequency and Monetary value (RFM).
This is a classic model for identifying your best customers, which assumes that customers who:
- Have made a purchase recently.
- Make regular or frequent purchases from you.
and:
- Spend a large amount with you.
are more likely to respond positively to future engagement activities and product offers.
Description
We use a simple, yet effective clustering algorithm based on the RFM model. The model assigns a score (1 – 10) to each variable, to identify top customers or lost customers.
Example of the RFM model
| Customer ID | R | F | M | Meaning |
| 11223344 | 10 | 10 | 10 | High recency, High frequency, High monetary → Top customer. |
| 22331144 | 1 | 8 | 8 | Low recency, High frequency, High monetary → Lost customer. |
You must provide data about the customer’s transactions as input. For example, the amount of money spent, the discount received and the purchase date for each transaction.
The output includes information and a score for each Recency, Frequency and Monetary variable.
Customer RFM clustering input data and output
| Input data | Output |
| Transactional data
Customer_id. Date of purchase. Order_id. Order_line_id. Amount. Discount. |
Recency. Frequency. Monetary. Recency score. Frequency score. Monetary score. |
Output examples
The output is provided as new fields that are associated with each customer profile.
Example of the output
| Customer ID | 11223344 |
| Recency | 2013-12-25 |
| Frequency | 12 |
| Monetary | €175.00 |
| Recency Score | 1 |
| Frequency Score | 6 |
| Monetary Score | 7 |
Tags:
- Clustering.
- Transactional.
Previous page: Customer purchase preferences| Next page: Customer Engagement