While the usage of speech interfaces goes back to the early days of speech recognition and Interactive Voice Response Systems of the 70s and 80s, the concept of a Conversational Voice User Interface (VUI) has been growing in popularity over the last decade. Large technical advancements have been made in VUIs over the past couple of decades, which has allowed conversational voice interaction to incorporated into mass consumer technology due to better technological capabilities. However, while commercial VUIs have been in the market for several years now, users still experience many inconsistencies with interaction and usability from VUIs - from recognition errors, to feedback, to error correction, etc. This leads to mismatched expectations and mental models, which can hinder the adoption of these interfaces in everyday life and prevent VUIs from reaching their full potential. I posit that this is due to the lack of focus on VUI designers – particularly, on support for new and existing VUI designers and on supporting for developing proper tools and resources for VUI design. This is reinforced by the lack of research on understanding the current VUI industry design space, and identification of the barriers and needs of VUI industry designers. In this thesis, I first contribute a large-scale understanding of the current VUI industry designer space, along with an identification of key barriers and needs that VUI industry designers currently experience, through surveys of professional designers and HCI curricula meta-reviews. Based on the findings from these surveys, I contribute a newly-developed synthesized set of heuristics representing the most common principles across VUI academic design literature and conducted an initial stage of consensus validation of said heuristics through a Delphi study with VUI industry designers.