The creation of market simulation models is increasingly widespread as it enables better enterprise planning and optimization of product delivery. In case of mass customization approach, which aims to simultaneously target scores of individual customers, understanding customer behavior under growing market stratification conditions is critical. This work describes development of a market simulator in which individual customers are represented as software agents, behavior of which is based on a set of predefined rules. Parameters in those rules can be either fixed or randomly follow some probabilistic distributions. Customers/agents operate in a market which offers them a line of products. These products have a set of customizable features, presumably addressing the needs of customer population. Our work describes in detail the decision process of an agent, market fragmentation design, and presents some sample results which can be used to determine desired variation offering of a product.