Metadata Analytics is a term used to describe a research field that utilizes quantitative methods and metadata for publications, patents, datasets, and other research entities to study science of science. Metadata analytics inherits the bibliometric and scientometric tradition while infusing novel data sources – metadata for datasets – to extend the traditional bibliometric and scientometric research. The large scale of metadata from scientific data repositories offers both opportunities and challenges in the quantitative study of science. This paper discusses the problems and opportunities that metadata analytics contends with from a methodological perspective. Using the authors’ experiences over the course of a multi-year metadata analytics project, the paper focuses on the subtle differences between methods and science (or means and end) that arise when conducting research in metadata analytics and, for the same reason, bibliometrics and scientometrics . Metadata analytics is both a methodology and a research field. The intertwining of methods and science in metadata analytics can create pitfalls for researchers. Steering clearly between the means and ends in metadata analytics is essential to produce good science.