top of page
2 days ago
3 min read

How should our fashion brand build RFM segments after returns and exchanges?

By Trayan Hristov · Founder, Thrivelia · 19 November 2026

A customer orders three sizes of one dress, keeps one and sends two back. If your VIP programme counts gross spend and line items, one outfit can buy them a lot of status. I’d give the CRM team a different starting file: one customer, one genuine shopping occasion and the value of what they kept.

Decide what recency, frequency and value actually mean

For recency, use days since the latest qualifying purchase occasion, with an explicit rule for returns that are still unresolved. For frequency, count separate paid shopping occasions. An exchange sent from the warehouse is linked to the original occasion. A later new full-price purchase is another occasion. For monetary value, use net merchandise value or contribution after returns under a consistent cost definition. Capture store orders where customer identity is verified and report the unmatched share.

Compare customers with similar purchasing opportunities. Recent full-price basics buyers, seasonal occasionwear shoppers and gift buyers may have different natural clocks. If the segments become too small, keep the main RFM bands simple and attach a product-family or gifting flag. Scoring 125 combinations is rarely helpful if they all receive the same campaign.

Follow three customers through the rules

Illustrative history

What gross reporting can suggest

Useful next action

Three dress sizes, one kept, first occasion

High spender

Fit support and care for the kept dress

Four kept basics orders across six months

Repeat full-price buyer

Check preferred size and relevant new colours

One high-value coat every winter

Inactive most of the year

Assess at the next outerwear buying window

Before assigning VIP status, let returns mature or mark the status provisional. Don’t penalise people for using a legitimate exchange policy or remove earned public benefits contrary to their terms. A private analytics segment and a published rewards tier are separate commercial decisions.

Backtest the segments before changing customer treatment

Freeze month-end labels using only the information available on that date. Compare next-90-day contribution and repeat occasions across segments. If a “high value” band is driven by one large order that is later refunded, examine how often this happens and revise the timing rule. If two adjacent bands behave similarly and receive the same offer, combine them.

Then test the action within the segment. For example, give some eligible full-price repeat buyers an optional restock preference and leave others on the current experience. Track kept orders and margin per eligible customer. A segment that predicts purchasing does not prove an offer to that segment earns any extra purchases.

The practical test is whether your new labels change a decision. Merchandising can see recurring size demand, service can see unresolved exchanges, and CRM can recognise a returning buyer without paying for an order they were going to place anyway. Apparel returns and exchanges · Shopify customer reports.

Make a segment move without losing its history

Keep a month-end snapshot. If an exchange closes in the middle of next month, update the live experience while preserving the original analytical label. Use the snapshot to build a migration table showing how many first-time buyers became second-time buyers, how many valuable active customers moved beyond their usual season, and how many shifts came from returns being processed. That separates real changes in customer behaviour from the ordinary timing of a warehouse.

Ask ChatGPT or Claude about your own numbers

Using these anonymous purchase occasions and return-adjusted values, build RFM bands for fashion. Flag unresolved returns, compare seasonal buying opportunities and specify one different action for each segment.

Related reading

 
 
 

Comments


bottom of page