Copying the data is the easy part
The thesis: on a large reporting platform, moving the data is rarely what overruns.
If you take one thing from this guide, take this: on a large reporting platform, moving the data is rarely the part that overruns.
The data has a known volume, a measurable throughput and a finish line you can see. What overruns is everything attached to it — the loading pipelines, the procedural code accumulated over a decade, the semantic layer, the four hundred reports of which nobody can say which forty matter, the entitlements model that exists half in the database and half in the BI tool, and the operational habits of the team that runs it at three in the morning.
I would plan on data movement being a minority of total effort. It is the part that gets estimated most carefully because it is the part that is easiest to estimate.
That has a direct consequence for method selection. Teams routinely optimise the migration method for data movement speed and then discover that the method has made the other work harder — a physical transport that preserved a physical model they wanted to change, or a lift-and-reshape that invalidated every piece of SQL in the reporting layer. Choose the method that leaves you with the least remaining work, not the method that moves bytes fastest.
Size matters, but it does not settle the choice. Complexity does — Complexity, not volume is why.