Here's a possible differentiation that no one's brought up: video-centric features.
The 70D has the new dual-pixel autofocus during video, and its inherited the All-I recording option found in the current-gen XD full frame cameras. But early sample footage suggests image quality isn't much better than that found in earlier ASP-C Canon models. Lots of artifacts, sub-HD resolution, etc.
That means the 7D II could feature some kind of significant step-up in video features. Maybe it will have a cleaner image, like the 5D Mark III (though I doubt it, if the 7D Mark II shares the 70D's sensor). Or maybe it will be a little sharper, like the 1DX, or even like the 35mm crop mode in the 1D-C. Or maybe Canon will do something really surprising, like implement decent video encoding. The much-ballyhooed All-I recording feature is better than the original implementation, but something that uses 4:2:2 color space would be nice, or that legitimately had a broadcast-ready bit rate.
Other rumors mentioned a 60fps burst mode, which has interesting photographic applications. That could produce 2.5 seconds of RAW 4K slow-mo footage, which opens up a world of applications currently unavailable to any Canon shooter--including the 1D-C crowd.
I think Canon could go this route without harming its other products.
The full frame models will have enough differentiation purely because of their larger sensors. The full frame aesthetic isn't possible on an ASP-C camera, at least not unless Speed Booster-style adaptors become more reliable and widely-adopted. For film people, it could make sense to have both an ASP-C body and a full-frame body, if Canon releases these hypothetical video enhancements. For stills people, meanwhile, if the 7D II shares the 70D's image sensor, full frame cameras are likewise still protected. So these video features present no loss to the XD full frame bodies, excluding losses that Canon is already evidently ready to accept (e.g. the company seems ready to accept some 5D Mark III losses among sports shooters by making the 7D Mark II a faster camera with a comparable autofocus system)
What about the C-series cameras? Well, the 7D II won't have the ergonomics of a C300, and that matters. If it didn't, Canon wouldn't be selling so many of the things. The RED Scarlet isn't much more money, and it produces better image quality in many ways. Still, the C300 still does very well because it allows faster workflows while offering, for most applications, video quality that's close enough to what one gets from RED. Plus there are other factors, such as built-in ND filters, audio ports, etc.
Also, Canon is going to have to move the lower-level C-series cameras to 4K relatively soon. Maybe not in the next year, but if the 7D Mark II is a 2014 camera, I wouldn't be surprised if a C300 mark II and a C100 mark II appeared shortly thereafter. If Black Magic ever gets its manufacturing act together, and if Sony keeps pushing hard on price, Canon's hand might be forced even sooner. So I don't think superior video specs on a 7D Mark II will cause any trouble with C-series bodies.
I constantly hear people talk about backlash when a company releases a superior spec on a downstream product-- i.e. is it possible some 5D Mark III users will be pissed if the 7D Mark II comes along with the video features they actually wanted? Sure, some people will grumble-- but they'd grumble more if Canon keeps relying on the same crappy video codecs.
The video improvements I've described could also help Canon with the Magic Lantern angle. I think Canon likes Magic Lantern in the sense that the hack encourages budget filmmakers to buy Canon DSLRs instead of, say, Nikon DSLRs, or something like the Sony FS100. But Canon also gets a lot of flack over Magic Lantern because the hacks illustrate how much Canon intentionally handicaps hardware. From Canon's perspective, this criticism isn't that big a deal, as it doesn't impact the markets where Canon makes its money-- e.g. consumers buying Rebels. But the markets it DOES annoy are still lucrative, both because there's a growing number of people in them, and because these people tend to buy more accessories than the average consumer. You'll find a million consumers who stick with the kit lens-- but try finding a wannabe filmmaker who hasn't begun investing in some fast primes, or who isn't salivating over something expensive, like the creamy bokeh of the 50L?
This is kind of a rambling post, but the point is this: I see lots of reasons for Canon to differentiate the 7D Mark II's video features in a major way, and not a lot of risk if the company chooses to do so.
