Okay, the best way to figure this out is to try Part A of both methods and time them. I'm betting this is where the problem is. Because you're creating the clustered index prior to data load, it's forcing the data to get reordered as it's inserting.

But that's a guess. As I said, the best way to know for sure is to ignore the other parts, and just do Part A of both methods in two separate SSMS windows. Time it, have it Display Execution Plan, and look at the plan for both bits when it's done running.