When the University of Michigan Press was surprised to observe how poorlytheir humanities publications were represented in their institution's Research Information Management system, they started to look for explanations. This investigation expanded to document, for the first time, all of the paths our metadata takes to reach various content distributors, discovery services, and otherlinks in the information supply chain, whether by automated or manual processes (or a combination of the two). The results revealed a tangle of criss-crossing, overlapping, and sometimes redundant metadata flows. This session will present a number of workflow diagrams to illustrate this status quo, with an emphasis on the prevalence or absence of persistent identifiers throughout the system, and identify a number of challenges that currently prevent our metadata from reaching its audience.