Customer segmentation

What customer segmentation delivers

Customer segmentation divides customers into groups with similar behaviour. The benefit lies not in the division but in what follows: anyone who knows a group buys differently can treat it differently. Without that split a shop treats everyone the same and the approach becomes arbitrary.

The difference from the target audience lies in timing. A target audience is the set of possible customers before the purchase, a segment is the actual customers after it. Segmentation therefore works with behaviour and not with assumptions.

By which criteria customers are grouped

Four criteria carry in retail. The first is value: revenue or contribution margin per customer over a period. The second is frequency: number of orders and the gap between them. The third is recency: when the customer last bought. The fourth is source: through which channel they arrived.

From value, frequency and recency a simple grid with four groups emerges. The best customers buy a lot and often. The loyal buy regularly with a smaller value. The occasional buy once. The lost have not bought for a long time but were active earlier.

Which segments carry in retail

• One time buyers: bought once and then never again. The largest group and the one with the greatest potential.
• Regular customers: steady purchases with a stable value. They carry the result and need care rather than advertising.
• Lost customers: were active earlier and have not bought for months. Recovery pays for only some of them.
• New customers with a high first value: few purchases but a high value. They are the most valuable group to develop.

How addressing follows from it

A different treatment follows from the division. Regular customers receive new content or products, not discounts. One time buyers receive a pointer tied to the first order, such as consumables. Lost customers receive one single attempt with a clear occasion. New customers with high value are addressed personally where volume allows.

What matters is that every segment gets its own figure. Anyone measuring success by a shared value does not recognise which treatment worked.

Where the division misleads

The first point is time. A customer buying twice in one month and then never again looks strong in that month and is lost afterwards. Segmentation over short periods leads to false conclusions.

The second point is segment size. A segment of twenty customers allows no statement, even when the values look clear. The third point is the cause. The division shows what happens, not why. A customer who does not return may have had a delivery problem, and that cannot be seen from the figures. The fourth point is the origin of the data: orders exist, on site behaviour may not, and both together give a different picture than orders alone.

Frequent questions

• How many segments? Three to five. More cannot be treated differently.
• Do I need a CRM? Not to start with. An evaluation from the shop system is enough.
• What is the most important criterion? Recency, meaning when the last purchase happened.