Sustainability as a guiding idea for societal and economic development causes a growing need for reliable sustainability assessments (SAs). In response, a plethora of increasingly sophisticated, standardizAed, and specialized approaches have emerged. However, little attention has been paid to how applications of SAs in different contexts navigate the challenges of selecting and customizing SA approaches for their research purposes. This paper provides an exploration of the context-specific conditions of SA through a case study of three research projects. Each case study explores the different approaches, methodologies, as well as difficulties and similarities that researchers face in "doing" SA based on the research question "What are common challenges that researchers are facing in using SA approaches?" Our case study comparison follows a most different approach for covering a wide range of SA applications and is structured along with three key challenges of doing SA: (i) Deliberation, learning and assessment; (ii) normative assessment principles; (iii) feasibility, especially regarding data quality/availability. Above all, the comparative case study underlines the role and importance of reflexivity and context: We argue that a more explicit and transparent discussion of these challenges could contribute to greater awareness, and thus, to improving the ability of researchers to transparently modify and customize generic SA methodologies to their research contexts. Our findings can help researchers to more critically appraise the differences between SA approaches, as well as their normative assumptions, and guide them to assemble their SA methodology in a reflexive and case-sensitive way.