I would agree that Sega's console future was salvageable in 1994, if not even later. The 32X came out in November 1994 so it could have been quietly yanked in late spring 1994 and the public would not even know about it. Heck, it could be yanked even after the CES announcement, with Sega saying something to the effect of "We're focusing on the Saturn as our future console." Sega would be damaged by this, but the damage would pale in comparison to what the 32X did. Instead of throwing major dollars at the 32X through the holiday season of '94 and first part of '95, don't even release the 32X. Stand behind the Genesis, and keep the CD going in the background. True, the CD had limited success, but it did have some great exclusive games. But the company's biggest asset was the Genesis. By late 1994, it had a very strong install base.

Here's a timeline that would likely give Sega a better future:

Let's start on May 9, 1994. Everything Sega's done prior to May 9, 1994 is done. The Genesis and Sega CD are their consoles. Game Gear is their portable system. They've even got the CDX out. SNES is just hitting its prime (which I consider to be the three full years 1994, 1995, 1996). 32X is in development but very few outside the business know: Summer CES 1994 hasn't happened yet.

By June 23, 1994: Cancel the 32X before CES. Don't announce it. Word will get out years later on retro game forums, but for '94 the secret is probably safe.

Remainder of 1994: Focus on the Genesis. Keep making Sega CD games in the background, but Genesis games are priority 1. SNES is coming out with expansion chips like the Super FX. Sega is falling behind technically, so it needs to go the expansion chip route for high end games. Sega's got the Sega Virtua Processor. Use it in several games over the 1994-1996 period. Get working on Sonic X-treme, but start its life as a Saturn game due out in '96.

1995: From January-August, keep focus on the Genesis. Announce "Saturnday" launch in March (they actually did this). Work with third parties and retailers before and during the September 2, 1995 launch. Go for at least 10 launch games, including a good VF2. Kill regional lockout. All Sega consoles from here on should be region free. Localize a majority of Saturn Japanese games. Some games should stay in Japan (i.e. dating sims, mahjong simulators, etc), but things like shooters and fighters need to be in the US. For the holiday season, push the Saturn, but maintain a focus on the Genesis as well, continuing to push out good titles. This is the last year that most focus goes to the Genesis. Start to shut down the Sega CD upon the Saturn's release.

1996: Push more Saturn games. Simplify the hardware and have the first price drop late in the year. Maintain Genesis support - this is the last year with a large number of Genesis releases. Overall, this year is mostly Saturn, but stand behind Genesis as a low cost option. With the Saturn out, this is the year to kill off the Sega CD for good. On Saturn, treat third parties well. This is also the year to release Sonic X-treme. It didn't make it in real life because it started on the Genesis, moved to the 32X, and finally was translated for the Saturn. Cut out even just the 32X bullshit, and Sonic X-treme can make it in stores in time for Christmas 1996. 3D Blast can be the Genesis Sonic swan song.

1997: DON'T SAY "THE SATURN IS NOT OUR FUTURE". Development of Dreamcast should be in earnest by now, and the buzz about the Dreamcast just starting. With third parties treated well, there should still be many cross-platform third party games that appear on the Saturn (And maybe a few exclusives). Keep the number and quality of first party Saturn games high. The 3D weaknesses of the system are becoming apparent, so emphasize some good 2D games as well. 2D wasn't totally dead by '97.