I am not sure about what are you asking exactly. I think the most important thing to decide on a methodological framework is the research question. The research design depends on what are you looking for.
To put it simply, as with all research, the researcher begins with a burning desire to answer a question about a problem. The researcher employs a map or theoretical framework to provide direction for the research. The researcher determines if the study is focused on single or multiple entities as well as will the study be in the participants' natural environment or context.
If you think of theory as a set of interconnected statements that link conditions, the social mechanisms they trigger, the conditions under which the mechanisms operate, and the effects these mechanisms produce, then it seems rather difficult to 'build' such a theory with any one project.
You can of course 'develop' such a theory - i.e. advance an existing theory - by deriving questions from it and answering them using qualitative methods. Unfortunately, most of these middle-range theories are not in a state that supports their development. You usually have to reconstruct the existing theory from bits and pieces in the published literature. Once you have done that, you have a framework in which you can ask a theoretical question that can be answered with qualitative empirical methods.
The most popular textbooks for the two basic versions of GT both include chapters on writing up the results: Constructing GT (Charmaz, 2014), and Classic GT (Holton & Walsh, 2016).
Beyond that, I would suggest finding examples in your own field by searching Google Scholar.