Skip to main content

S4. Persona Matching + Clustering (Uni-President)

Published 2026-05-14Updated 2026-06-301 min read

5 Lifestyle Personas

PersonaSignal
Daily CVS ShopperDaily 7ELE + beverages · bento
Weekend Family Hyper ShopperWeekend Carrefour + groceries · household goods
Café MorningWeekday-morning Starbucks
Multi-BU OmnishopperActive across 7ELE + Carrefour + Starbucks
Seasonal TrendsetterBuys new SKUs immediately + active on social

Data Mix

  • RFM
  • BU affinity (CVS / Hyper / Cafe share)
  • BUTransfer frequency
  • Social personas (Dcard · Xiaohongshu (小紅書) keywords)

Output UI

  • 5 persona cards
  • KMeans 6 clusters (LLM-labeled)
  • PCA scatter plot

Demo Scenarios

  1. Single-record match → "Multi-BU Omnishopper 0.65"
  2. Bulk 5K → 6 clusters ("Morning Starbucks + Lunch 7ELE office worker", etc.)
  3. Cluster → cross-BU campaign (S5)