What is RFM Analysis?

RFM analysis is a classic segmentation method scoring customers in three dimensions: Recency, Frequency, Monetary. Score 1-5 in each; the resulting three-digit code (e.g., 555 = best, 111 = worst) places each customer in a meaningful segment.

// the three dimensions

• R (Recency) — time since last order. 1 = long gone, 5 = recent purchase.
• F (Frequency) — orders in window. 1 = rare, 5 = frequent.
• M (Monetary) — total spend. 1 = low, 5 = high.

// scoring method (quintile)

1. Sort customers by R; top 20% gets 5, bottom 20% gets 1.
2. Sort by F similarly, score 1-5.
3. Sort by M similarly, score 1-5.
4. Combine into 3-digit RFM code per customer.

// segment map

• 555, 554, 545: Champions / VIPs — most valuable, most loyal.
• 5XX (high R): New / active customers.
• X4X-X5X high F, low R: At-risk loyal — proactive campaign.
• 11X, 12X (low R, low F): Lost — winback, then archive.
• 551, 541: New low-spend — upsell opportunity.
• 515, 514: One-big-purchase — recovery focus.

// where RFM works

• E-commerce (always).
• Subscription / box services.
• Restaurant chains (with loyalty data).
• Banks / finance (transaction-based).
• SaaS partially — adapt with usage frequency + plan size.

// modern RFM (RFM+)

• RFML — adds Length (relationship duration).
• RFMP — adds Profit (margin not just revenue).
• RFM + Predictive — adds future-purchase probability (CLV models).

// after segmentation, then what

1. Calculate funnel and LTV per segment.
2. Prioritize high-value + at-risk segments.
3. Differentiate message, channel, offer.
4. Track segment movement monthly (who moved from VIP to At-risk).