RFM segmentation for e-commerce: a practical guide

Recency, frequency, monetary value

The RFM method ranks each customer on three criteria drawn from their order history:

• Recency: how many days since their last order. It is the most predictive criterion: a recent customer is far more likely to order again.
• Frequency: how many orders they have placed.
• Monetary value: how much they have spent in total.

These three numbers already exist in your store, on PrestaShop as on Shopify: you only need to pull them from your orders.

Computing the scores

The most common approach gives each customer a score from 1 to 5 on each criterion, by splitting your customers into five equal-sized groups (quintiles):

1. Export your customer list with the date of their last order, their number of orders and their total spend.
2. Sort on each criterion and give 5 to the top 20%, 1 to the bottom 20%. For recency, the most recent get 5.
3. Combine the three scores: a 5-5-5 customer ordered recently, often and a lot.

Leave out cancelled and refunded orders, and pick a window consistent with your buying cycle: a customer who buys a mattress every eight years is not “lost” after six months.

Six segments to act on

The 125 possible combinations group into a few readable segments:

• VIP: recent, frequent, high spend. The core of your revenue.
• Loyal: order regularly, without necessarily spending a lot.
• Potential: recent, a few orders, promising.
• New: a recent first order.
• At risk: good customers who have not been back for a long time.
• Lost: no order for a very long time, little history.

What to do with each segment

• VIP: no systematic discounts, they already buy. Early access, personal attention, a loyalty programme.
• Loyal: complementary products and upgrades to grow their basket.
• Potential: guide them to the second or third order, the one that builds the habit.
• New: a welcome sequence, and a concrete reason to come back in the following weeks.
• At risk: the most profitable segment to work on. A personal reminder at the right time costs far less than acquiring a new customer.
• Lost: one reactivation attempt, then take them out of regular sends to protect your deliverability.

Spotting customers about to churn

A segment is a snapshot. What matters is movement: a VIP customer sliding to “at risk” is a warning, far more than a customer who is already lost. Compare each customer with their own pace: someone who orders every month and has not been back for two months sends a strong signal; someone who orders twice a year does not.

To prioritise reminders, put a number on what is at stake: what those customers usually spent. It is the cost of inaction applied to customers.

Up-to-date segments without exports

Recomputing scores every month from exports quickly gets tedious, and segments that are three months old miss precisely the customers who are leaving. Lumexio segments your customers automatically with RFM, estimates each one’s churn probability, flags customers who change segment and suggests costed win-back actions. Available for PrestaShop and Shopify.