Part 1 of this document describes the background, format, and early groundwork that went into the development of a test sponsored entirely by private enterprise. The discipline imposed by a financial bottom line imposes special pressures but also offers new opportunities. This private enterprise model is a multi-constructional process where language testing research and literature; English-as-a-Second-Language teacher-cum-test developers; a transient pool of teachers; non-language specialists, directors, managers, and accountants; and the ever-present need for performance-reliable second language assessment all play a role. Close interaction between researchers and practitioners is essential. Part 2 explains how opportunities to use data from such a test-in-development led to more substantial and valid research designs that could be executed in-development to improve assessor training and reporting descriptors. There is a need to assess the impact of varied test formats and task features on interaction and to explore how these events vary in different cultural settings. The design of this test was an attempt to maximize the conversational style of interaction through features of computer-assisted test design, including a succession of holistic evaluations incrementing to a final score, reducing the processing and memory burden on the native speaking interlocutor, and positioning the evaluations at junctures that could allow the native speaking interlocutor to avoid sudden pre-closure and closure moves once the assessor has gotten what is needed from the student. Minimizing topic hopping is essential to make the assessment more like a conversation than an interview. (KFT)