I have always enjoyed building systems retro-commissioning because it’s like a super-sized lab experiment. The process boils down to collecting, reviewing, and understanding available data and applying the laws of physics to fill in the missing pieces. This often requires testing numerous theories until one of them is proven to answer an outstanding question about how a system functions.
Before responsibly making performance and energy conservation recommendations, it’s necessary to understand the existing systems’ operation to a level of detail often not documented in as-built drawings or manuals. As such, retro-commissioning conclusions and recommendations should never be based solely on available documentation. That would be like limiting a research project to the literature search. Documentation review is the necessary first step, but it should only be used to formulate questions and theories to be investigated in the field. Data analysis and experimentation is always necessary.