Breakthrough innovation inside established companies—innovation that creates new revenue streams—often requires business model innovation in addition to innovation in product or offering. Unfortunately, new business models are rife with uncertainty. They may require new channels or new partnerships, appeal to new customer sets, rely on different revenue models, or require unfamiliar technologies. The risk associated with creating new business models is a major reason high-potential new businesses are not launched inside established companies. Well-designed business experiments are a valuable tool for reducing that risk. Although business experiments are not new, much of the literature concerning such experiments focuses on software and Internet businesses, where experiments are easier to conduct, and on startup companies rather than large corporations. This paper addresses the conduct of experiments in companies that produce physical goods and addresses the issues associated with conducting experiments within an established company. It describes methods used to design experiments, to focus them on critical issues, and to manage them in a corporate context.