Why segment an email base

A mass send to the whole base is fast but ineffective. A premium-brand buyer receives a message
about a 70% off sale and unsubscribes. A newcomer with no purchases receives thank you for your
loyalty and is baffled. The result: a damaged domain reputation, more spam complaints, a falling
sender score.

Email segmentation solves this by splitting the base into groups with different content. The same
send date, different messages. The outcome is a higher open rate, fewer unsubscribes and more
attributed revenue per message sent.

The main types of segmentation

Type Signals Typical scenario
Demographic Gender, age, city Different product selections
Behavioural Purchases, views, clicks Recommendations based on history
Lifecycle New / active / dormant / VIP Onboarding, reactivation
RFM Recency, frequency, monetary Differentiated offers
Category Affinity for catalog sections Specialist selections

RFM — the foundation of e-commerce segmentation

RFM segmentation splits the base along three dimensions:

  • Recency — how long ago the last purchase was
  • Frequency — how often they buy
  • Monetary — how much they have spent
VIP segment:       R=1, F=5, M=5  -> priority, exclusive access
At-risk segment:   R=4, F=4, M=4  -> reactivation with an offer
Dormant segment:   R=5, F=2, M=2  -> win-back campaign or removal

Tip: dormant users who have not opened a message for 6+ months are better moved into a
separate reactivation branch with a single message. No reaction — remove them from the active
base. That lifts deliverability and sender score for every other segment.

Common mistakes

  • Too many segments. Twenty micro-segments on a base of 10,000 means samples too small for
    statistically significant tests and a content production load that is too heavy. Start with 3–5.
  • No exit from a segment. A user who has stopped being dormant keeps receiving reactivation
    messages although they have already bought. Use dynamic segments with an exit condition.
  • Segmentation without content personalization. Five segments created, identical messages sent.
    The minimum personalization is the subject line and the recommendation block.