Templefield Technologies
Data Science

Machine Learning-Driven Customer Segmentation

Machine learning-driven estimation of customer value using XGBoost

Industry
Healthcare / Diagnostics
Client
Unilabs Switzerland
Engagement period
July 2024 – September 2024
Delivery mode
Remote
ML-driven customer segmentation by value

Situation

The client needed a better understanding of customer value, market potential, and regional market share. Fragmented data across systems prevented effective decision-making and held back strategic growth initiatives.

Objective

Templefield Technologies was engaged to develop a data-driven methodology that consolidates customer information, establishes clear value criteria, and applies machine learning to estimate customer value and drive segmentation.

Approach

  • Integrated more than 10 external data sources with internal data using SQL
  • Defined standardized, business-aligned value principles for scoring customers
  • Trained and evaluated an XGBoost model for customer value estimation
  • Identified promising new customers and expansion opportunities through ML
  • Established data governance frameworks to sustain long-term data quality

Results

The engagement delivered actionable customer segmentation, a clear assessment of market potential, and the discovery of profitable expansion opportunities. The accompanying data-governance framework maintained quality standards, supporting targeted marketing and sustained business growth. The resulting initiatives increased conversion by 11% and upselling by 14%.

Technology stack

XGBoostSQLPythonPandasMicrosoft Azure Data LakeGoogle Places APIData ScrapingREST APIPower BIPostman
All work