The preceding chapter covered all the preparatory actions that should be taken to minimize the risks associated with an experimentation phase of a project. These range from conducting research that informs the options available for solving the problem to building useful functions that the team members can leverage during the prototyping phase. We will continue the previous scenario throughout this chapter, a time-series modeling project for airport passenger demand forecasting, while focusing on methodologies to be applied to experimental testing that will serve to reduce the chances of project failure.
We will spend time covering testing methodologies simply because this stage of project development is absolutely crucial for two primary reasons. First, at one extreme, if not enough approaches are tested (evaluated critically and objectively), the chosen approach may be insufficient to solve the actual problem. At the other extreme, testing too many options to too great a depth can result in an experimental prototyping phase that risks taking too long in the eyes of the business.