HeyMark
Content

Blog

RFM analysis: meaning and customer segmentation example

HeyMark3 min read
Three groups of wooden tokens in front of a bar on a tabletop.

RFM analysis summarizes each customer’s purchase history with three measures: recency (days since the last purchase), frequency (number of purchases) and monetary value (amount spent). IBM SPSS documents these variables and their combined scoring; it does not set universal thresholds or prove which campaign will work. IBM SPSS, RFM Analysis

A reproducible example with three customers

Set a cutoff date, October 29, 2026, and the same window for everyone: the previous 180 days. Count completed purchase transactions, not line items: an order with several products is one purchase. Exclude cancellations and count partial returns in monetary value by subtracting the amount refunded, so spend is net of returns. Suppose the results are:

Fictional customer Days since last purchase Purchases in 180 days Net spend in 180 days
Ana 10 4 $240
Bruno 40 2 $100
Carla 90 1 $30

For a simple demonstration, set cutoffs before interpreting the data: recency of 0–30, 31–60 and 61+ days; frequency of 1, 2 and 3+ purchases; spend under $50, from $50 to less than $150, and $150+. Ana falls into recent, high-frequency and high-spend bands; Bruno lands in the middle bands; Carla falls into the lower activity and spend bands. Those labels describe only this window and these rules.

Do not calculate quintiles from three observations: there are too few customers for five useful groups, and ties can produce repeated cutoffs. With a large dataset, quintiles show relative positions within that dataset; they shift with its composition and are not natural thresholds. IBM explains R, F and M scoring, but the cutoffs here are illustrative, not IBM recommendations.

Turn a label into a cautious decision

For recency scores, fewer days since the last purchase means a higher score; more days means a lower one. If Carla has not purchased in 90 days, first check the product’s normal buying cycle. A pause may deserve attention for a replenishment item; it may be normal for something that lasts years. Check margin and seasonality before choosing whether to contact her. The segment does not show that Carla will respond or that Ana is profitable or a “VIP.”

RFM describes past purchases. It does not measure cohort retention on its own or explain why someone stopped buying. To study customer retention over time, see the customer retention rate guide. Before running a campaign, verify contact permissions and evaluate results against an explicit time window and comparison group.

Plan your next post in HeyMark.

Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.