There is an occasional objection/concern I hear with the idea that in a 2/1 sequence, trumps agreed at the two-level, cuebidding starts immediately. Namely, there is some concern that pattern bidding might help on some hands and might be preferable.
Here's how I think about this, in a slightly different take than my usual response.
2/1 sequences typically start with some pattern development (we announce a few suits before agreeing trumps) and then convert into pure cue, albeit with some pattern possibilities (picture bids).
Jacoby 2NT sequences start with a premise that pattern is key, with cuebidding available only after pattern is developed. (Hence why my own structure is as it is -- very pattern specific.)
I think this makes a lot of sense, personally, so long as you develop the Jacoby 2NT structure beyond the simplified versions usually used. For a simple example, the methods I use (from Ken Eichenbaum) enable description of which stiff is held when Opener has a 5-5 two-suiter, a very pattern-oriented move.
This obviously means, of course, that your pattern development is somewhat hampered, because the 2NT call takes up so much pattern development space (you don't get to start the pattern development with a few other-suit calls before reaching 2NT, as in the 2/1 sequences). So,if you wanted to maximize pattern development, perhaps reversing these would make sense.
However, I think standard theory has pattern first in both sequences, and then cuebidding, which results in too much pattern development and way too little in the way of cuebidding.
Plus, in my opinion, purity as to cuebidding is more conditional, meaning that the most effective cuebidding structure requires, to some degree, pre-work in the form of the few bids leading up to fit establishment.
It still boils down, I suppose, to focus and preference and what you are comfortable with.