The Customer Segment Nobody Wanted to Talk About.

2026-02-17 · Source: Data Science on Medium · Field: Business & Management — Marketing, Branding & Advertising, E-commerce & Digital Commerce, Operations & Process Management · Depth: Intermediate, long

Summary

An analysis of a publicly available retail dataset, comprising over a million transactions, challenges the conventional wisdom of focusing solely on "VIP customers." The study involved extensive data cleaning, reducing 1,067,371 rows to 805,549 clean transactions by removing anonymous purchases, returns, cancellations, and zero/negative priced items. Using RFM (Recency, Frequency, Monetary) analysis combined with K-means clustering, the data revealed five distinct customer segments: Mega Whales (4 customers, >10% revenue), VIP Whales (23 customers, >$100K spend), Champions (341 customers, ~$14K spend), Loyal Regulars (2,883 customers, ~$2K spend), and At-Risk (1,010 customers, inactive for 8+ months). Surprisingly, financial modeling projected the highest ROI (11.7x) from the "Loyal Regulars" segment, compared to 7.0x for VIP Whales, due to their large volume and potential for incremental growth.

Key takeaway

For marketing professionals and data scientists evaluating customer segmentation strategies, you should critically re-evaluate the ROI potential of your "middle-tier" customers. Your biggest growth opportunities might reside in these often-ignored segments, which offer significant leverage through volume and untapped potential, rather than solely focusing on already-maxed-out VIPs. Consider running a similar RFM and K-means analysis on your own data to uncover hidden value.

Key insights

Challenging assumptions with data can reveal overlooked customer segments with higher ROI potential than traditional VIPs.

Principles

• Real-world data is inherently messy and requires significant cleaning.
• RFM analysis combined with K-means clustering identifies natural customer segments.
• Highest revenue contribution does not always equate to highest marketing ROI.

Method

The method involves cleaning raw transaction data, calculating RFM scores for each customer, applying K-means clustering to identify segments, profiling each segment, and then modeling the projected ROI for targeted marketing strategies.

In practice

• Prioritize data cleaning before any segmentation analysis.
• Use K-means clustering for unbiased customer segmentation.
• Model ROI for each segment to identify true growth opportunities.

Topics

• Customer Segmentation
• RFM Analysis
• K-means Clustering
• Marketing ROI
• Data Cleaning

Code references

• danialakhan007/RFM-Customer-Segmentation-Analysis

Best for: Data Scientist, Business Analyst, Marketing Professional

Related on AIssential

• I Built a Demand Forecasting Model at a Real FMCG Company. It Failed First. (Part 1) — Segmenting diverse demand patterns before forecasting significantly improves model accuracy.
• What is customer segmentation? — Effective customer segmentation requires measurable, accessible, substantial, differentiable, and actionable groups, continuously updated and built on unified, high-quality data.
• Product Mix Emerged as the Primary Driver of Profitability — Profitability is driven more by product mix and margin health than by sheer sales volume.
• KKR backs UK’s newest unicorn in $80m round — Fresha's unicorn valuation highlights investor confidence in profitable vertical AI software platforms for the beauty and wellness sector.
• When Customers Churn at Renewal: Was It the Price or the Project? — Accurate churn attribution at renewal requires distinguishing between price shock, value exhaustion, and their interaction.
• PayPal Ads Delivers Bottom-Line Growth Advertisers Can Measure — PayPal Ads leverages cross-merchant transaction data for measurable advertising outcomes beyond impressions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Data Science on Medium.