There's an old proverb that goes:
For want of a nail the shoe was lost.
For want of a shoe the horse was lost.
For want of a horse the rider was lost.
For want of a rider the message was lost.
For want of a message the battle was lost.
For want of a battle the kingdom was lost.
And all for the want of a horseshoe nail.
The 32X is kind of like Sega's nail, or lack thereof. Obviously, it's not the only reason they had trouble, just as the kingdom in the proverb probably had more issues than one horse with a bad shoe. However, if I could go back in time and reverse just one of Sega's decisions, this would be it. No 32X means the Saturn probably gets off to a better start, especially in the US, and you can take that right on out to the Dreamcast putting up a real fight against the Playstation 2 however you like to imagine it.
An add-on like the 32X seems like such a neat idea in so many ways. I get why people only wish it had been done differently, rather than done away with altogether. However, it really should never have made it off the drawing board. It's simply a matter of it being foolish to divide your consumers and your development resources. There was no way you could successfully turn the Genesis into a next-gen system and simultaneously launch another stand-alone next-gen system, period. Since the Genesis makes a lousy backbone for a next-gen system, Sega's choice should have been pretty obvious.
My two-cent armchair retrospect strategy for Sega in 1994 would be to can the 32X and the Saturn both. Sony's Playstation was just too well-designed to take on directly with anything, let alone the mess that the Saturn was. Instead, drop the price of the Genesis and the Sega CD as low as possible to ride out the storm until 1996, then release a system with the processing power of the N64 and the CD storage of the Playstation. And have a damn 3D Sonic game ready at the release.