This document discusses performance evaluation of clustering algorithms for analyzing sales data of a steel company. It analyzes annual sales data using clustering techniques like K-Means and EM to group sales by products, customers, quantities and reveal patterns. The study finds that partition methods like K-Means and EM are better suited than hierarchical or density-based methods for analyzing the company's sales data. It also discusses various sales analysis dependent issues like product attributes, price fixation, new product development, and key performance indicators that were considered in the clustering analysis.