People often find it challenging coming into a poorly defined process. Walking through a poorly defined process requires rigorous work to re-define and test changes.
Git is a wonderful tool for that, but one thing that Git really doesn't like to do is store large datasets.
It can be challenging to use if you're not familiar, but the thoughts behind 'track the change as it flows along the codebase' doesn't just work for code development, but it can also work for process development.
You can start a project knowing nothing and simply begin storing information; as you learn more you can begin adding useful metadata, tracking additional components (or fewer!), and making the entire process much more rigorous.
Semper Gumbi!