.. _dpa-recency-frequency-and-monetary-value-rfm: ======================================================================== Recency, Frequency, and Monetary Value (RFM) ======================================================================== **Stage 4 · 🧩 Sampling, Partitioning & Segmentation** · Lesson 28 of 56 · *intermediate* :doc:`◀ Previous · Clustering <27-clustering>` · :doc:`Next · RFM Analysis ▶ <29-rfm-analysis>` · :doc:`↑ Section ` .. important:: **✨ AI-generated content.** This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it. Three questions about a customer ---------------------------------- How valuable is a customer? **RFM** answers with three simple questions, each read straight from transaction history: **how recently** did they buy, **how often** do they buy, and **how much** do they spend? These three numbers — **Recency, Frequency and Monetary value** — summarise a customer's behaviour compactly enough to rank an entire database. The three dimensions ---------------------- Each dimension is one measurement per customer: * **Recency** — days since their **last** purchase. Fewer days is better: recent buyers are far likelier to buy again. * **Frequency** — the **number** of purchases in a chosen window. More is better: repeat buying signals habit and loyalty. * **Monetary** — the **total spend** over that window. More is better: it captures the customer's economic value. Why all three --------------- No single dimension tells the whole story. A **big spender** who has not bought in two years is a **churn risk**, not a star; a **frequent** buyer with tiny orders is loyal but low-margin. Combined, the three give a **holistic** view that any one alone would distort. (Of the three, **recency** tends to predict future behaviour best, and monetary least.) Simple and proven ------------------- RFM's great virtue is **simplicity**. It needs only data every business already has — an order history — and no elaborate modelling. The technique dates back to **direct-mail** marketing in the last century, and it endures because it works: it reliably surfaces the roughly 20% of customers who drive most of the revenue. The next lesson turns these three raw numbers into scores. .. hint:: **Related lessons:** :doc:`RFM Analysis <29-rfm-analysis>` · :doc:`Creating Segments of Observations for Business Reasons (RFM) <30-creating-segments-of-observations-for-business-reasons-rfm>` · :doc:`Cross-Selling <22-cross-selling>` · :doc:`Clustering <27-clustering>` .. seealso:: **Source article** Adapted (context, re-expressed) in our own words from: `https://insightful-data-lab.com/2026/01/14/recency-frequency-and-monetary-value-rfm/ `__ (insightful-data-lab.com). .. tags:: purpose: reference, topic: data analysis, topic: data preparation, level: intermediate