Work

Henkel Budget Optimization

Case Study 2 December 2024
Machine Learning
Geo Analysis
Marketing
Business Strategy

A budget allocation case study where machine learning and geospatial clustering helped Henkel decide where every euro would have the most impact.

A red image of Henkel logo

Description

In this project, we collaborated with Henkel, a global leader in the DIY tools and consumer goods market. The objective was to optimize budget allocation strategies across different stores to maximize returns and operational efficiency. Together with Jonathan Bouniol, we conducted advanced data-driven analyses and provided actionable recommendations tailored to Henkel’s diverse retail network. The project integrated machine learning models and geospatial analyses to ensure a precise and impactful distribution of resources.

Key Features

  • Machine Learning Implementation: We utilized machine learning algorithms, such as linear regression and K-means clustering, to build predictive models for sales and identify patterns in budget allocations. Multiple scenarios were developed to demonstrate the impact of reallocating budgets on key performance indicators.

  • Data Analysis: Through comprehensive data exploration, we analyzed correlations between variables like budget size, sales performance, and regional factors. This allowed us to identify discrimination factors, enabling effective store clustering and informed recommendations for resource reallocation.

  • Marketing and Finance: Our strategy included A/B testing frameworks to evaluate the effectiveness of budget allocation changes, financial projections grounded in ROI metrics, and detailed analysis of KPIs such as customer acquisition cost (CAC) and average revenue per user (ARPU).

Technologies Used

  • Python: Core language for data processing and model building.
  • Scikit-learn: Linear regression and K-means clustering for budget optimization.
  • Pandas: Data cleaning, transformation, and in-depth exploration.
  • NumPy: Numerical computations and feature engineering.
  • Excel: Financial ratio analysis and ROI evaluation.

My Role

I partnered with Jonathan Bouniol on this Henkel Business Deep Dive. My focus was on the machine-learning side: sales forecasting, store clustering, and feature engineering. I also helped build the financial scenarios and A/B testing framework that turned our model outputs into actionable budget recommendations.

What I Learned

A model score is not a business decision. Henkel pushed us to translate every clustering result into a budget movement with a clear ROI narrative. This experience taught me to always ask: “So what?” — and to answer before the stakeholder does.

Outcome

The project successfully provided Henkel with a set of actionable strategies to optimize its budget allocation across stores. By implementing our recommendations, Henkel gained deeper insights into which stores and regions required additional funding and which could operate efficiently with reduced budgets. Our analysis highlighted an average improvement of 15% in ROI for test cases, and the clustering approach revealed untapped market potential in specific geolocations. The tools and models developed in this project are scalable and can be adapted for future financial planning and strategy development across Henkel’s global operations.