Despite significant research into authoring tools for interactive narratives and a number of established authoring platforms, there is still a lack of understanding around the authoring process itself, and the challenges that authors face when writing hypertext and other forms of interactive narratives. This has led to a monolithic view of authoring, which has hindered tool design, resulting in tools that can lack focus, or ignore important parts of the creative process. In order to understand how authors practise writing, we conducted semi-structured interviews with 20 interactive narrative authors. Using a qualitative analysis, we coded their comments to identify both processes and challenges, and then mapped these against each other in order to understand where issues occurred during the authoring process. In our previous work we were able to gather together a set of authoring steps that were relevant to interactive narratives through a review of the academic literature. Those steps were: Training/Support, Planning, Visualising/Structuring, Writing, Editing, and Compiling/Testing. In this work we discovered two additional authoring steps, Ideation and Publishing that had not been previously identified in our reviews of the academic literature-as these are practical concerns of authors that are invisible to researchers. For challenges we identified 18 codes under 5 themes, falling into 3 phases of development: Preproduction, where issues fall under User/Tool Misalignment and Documentation; Production, adding issues under Complexity and Programming Environment; and Post-production, replacing previous issues with longer term issues related to the narrative's Lifecycle. Our work shows that the authoring problem goes beyond the technical difficulties of using a system, rather it is rooted in the common misalignment between the authors' expectations and the tools capabilities, the fundamental tension between expressivity and complexity, and the invisibility of the edges of the process to researchers and tool builders. Our work suggests that a less monolithic view of authoring would allow designers to create more focused tools and address issues specifically at the places in which they occur.