This paper presents an experience gained in the developmentof a large component based application that handles massive federal and state government labor market information data. The process of transforming user requirements to deliverable software components are shown in various stages of development. Components developed are rated against existing specifications. Further, these components are classified based on their functionalities and location. Statistical information is presented for components, component types and re-use types. Our experience suggests two practical componentization methodologies and JavaBeansextensions. As per industry-taste, these extensions can become a part of future specifications in order to enable the development of fully component based applications. Wealso demonstrate uses of patterns to document our architecture.