Theories are, in the technical sense, our inter-related ideas and propositions about aspects of the world with which we conditionally (according to our inter-related ideas and propositions) explain those aspects [is that clear enough?].
Thus, theory-building is building those inter-related ideas and propositions.
However, in the sciences, we seek to test theory, and build ongoing knowledge from theories which pass empirical testing (modifying or dumping those which don't).
This level of theory which is amenable to testing consists of elements ( your inter-related ideas and propositions) which can be translated into the measurable and observable; in social science, Merton called these theories of the middle-range------because they were neither too abstract nor too focused, in which cases, such 'theories' would not explain very usefully.
Thus, as I see it, theory-building is building-up your inter-related ideas and propositions about aspects of your field of study, which can be made amenable to empirical testing, towards giving conditional explanations about those aspects of your field of study.
This has been a bit 'long-winded', but it may help.