What is recency-frequency-monetary (RFM) scoring?

Event-based segmentation groups people by specific actions they've taken. Instead of asking "who is this person?" it asks "what has this person done?"

Events are discrete actions with timestamps:

Account events: Signed up, verified email, completed profile, upgraded plan, downgraded, churned.

Purchase events: Added to cart, started checkout, completed purchase, requested refund.

Engagement events: Opened email, clicked link, watched video, downloaded resource, attended webinar.

Product events: Logged in, used feature X, hit a milestone, encountered an error.

Segmentation uses event combinations:

"Signed up but never logged in" identifies onboarding failures.
"Added to cart but didn't purchase" identifies abandoned cart opportunities.
"Used feature X more than 5 times" identifies power users.

Event-based segments are inherently dynamic. Someone enters the "abandoned cart" segment when they abandon. They leave when they complete purchase or the window expires. The segment membership changes constantly based on real behavior.

This approach requires event tracking infrastructure. But it creates the most actionable, timely segments possible.