Loyalty Register glossary
Segmentation
Dividing members into groups that share traits or behavior, so a program can give each group different offers, messages or rules.
Customer segmentation is dividing customers or loyalty members into groups that share traits or behavior, so the program can treat each group differently: different offers, messages, earn rules or service. Every program that sends more than one kind of message to its base is already segmenting, deliberately or not. (In biology, computing or image analysis, “segmentation” means something else; this page covers customers.)
Updated · By Loyalty Register Editorial TeamRelated terms
What are the four types of customer segmentation?
The four classic types are:
- Demographic: age, gender, household or income.
- Geographic: country, region, city or distance to a store.
- Psychographic: values, interests and lifestyle, often from surveys or zero-party data.
- Behavioral: what members do, such as purchase frequency, spend, categories, channels and redemption habits.
Many teams add value-based segmentation (current or predicted value, often CLV) and needs-based segmentation (what members say they want).
Which segments matter most in a loyalty program?
Behavior usually predicts behavior better than demographics, and a loyalty program records behavior other teams cannot see. Useful loyalty segments include:
| Segment | Built from | Typical action |
|---|---|---|
| Lifecycle stage | New, active, at risk, lapsed | Onboarding, nudges, reactivation |
| RFM group | Recency, frequency, monetary value | Rewards and offers by value |
| Points position | Balance close to a reward, or expiring | Redemption reminders |
| Tier distance | Spend or activity left to the next tier | Progress messages |
| Redeemer type | Never redeems, saves, redeems often | Catalog and messaging changes |
RFM analysis is the usual starting point. Tiers are segmentation too, with published rules members can see.
How many segments should you have?
As many as you can act on and measure. A segment is only useful if it gets a different treatment and you can tell whether that treatment worked. Ten segments that all receive the same generic email are no better than one. Many teams start with four to eight, then split into micro-segments once they have the content, offers and testing capacity.
What should you watch?
Segment membership should update as behavior changes; a label assigned last year will send the wrong offers this year. Each segment needs a control group if you want to know whether its treatment changed anything. Keep segments explainable, especially when they decide prices or benefits, because members compare notes. And check that the data you use is covered by your privacy notice and any consent you collected.