Power BI has a big issue when trying to load Excel files as a data source. The files could be very small (only a couple of megabytes), but if you do the wrong combination of steps (such as an append), the load is extremely slow. Even a straight load is very slow, for the size of the file.
I suggest that Power BI should load the Excel files once into memory, in Power BI's native table format (or whatever in memory), and then refer to that for all future steps.
Currently it will re-read the same file(s) from disk, many, many times - depending on the steps performed. I'm unsure why this overhead is needed.
If I was to perform the equivilant load and data maniputation using Python Pandas, it would load the excel files once, perform all the data manipulation, and be done in a fraction of the time.
Microsoft really need to pick up the game on this. I have no idea why they treat an excel file like a database. I lose hours of time whenever I deal with excel files, because of this. The worst is when you have a folder of excel files.