All case files
E-commercesegmentation1 weekPlaceholder

Customer segmentation for a small Shopee fashion seller

Owner was sending the same promo voucher to every customer. We split the base into 4 groups and rewrote the campaign to match.

Client · Solo Shopee fashion seller, ~600 customers


A /

The problem

  • Owner ran a monthly broadcast voucher to every past customer. Conversion was under 2% and reply rate was even lower.
  • No clarity on which customers were one-off buyers versus genuine regulars. Big spenders treated the same as bargain hunters.
B /

The approach

  1. 01Pulled 18 months of Shopee order export into Python and built RFM features (recency, frequency, monetary value).
  2. 02Ran K-Means clustering, validated the 4-segment split with the owner against real customers she could name.
  3. 03Wrote a short playbook with one tailored message and offer per segment.
C /

What was delivered

  • Excel file with every customer tagged into one of 4 segments
  • 1-page playbook: which segment to message when and what voucher value to use
  • Walkthrough call (45 minutes) covering how to redo this in 6 months

02Results

What changedafter.

  • 4segments

    VIP, Regular, One-off, At-risk

  • 2.1x

    voucher redemption rate on the next campaign vs previous month

  • RM 0

    spent on customers who hadn't responded in 12+ months (was being broadcast to all)

NoteFrom the desk
Placeholder testimonial. Real quote landing soon as the first wave of cases wraps up.
BroData

Want something similar?

Send the brief. WhatsApp reply within the day. If it makes sense for your business, we go from there.