Break down your Analysis requirements into measures, dimensions and descriptors, essentially translating the requirements from natural language into “data-speak”. Consider relationships between different elements of your data and how these will impact both the analysis of the data and the underlying structure of the data model.
Gather the data that will power the analysis and help inform your decision. At this cellular level, look for specific fields within specific data sets. Is the raw data complete, up-to-date and accurate? Since your decision will ultimately be informed by your data, data quality is of the utmost importance.