Unlike the variable roster sizes of certain college teams (hi, Houston Nutt! How'd you find this on a computer? Someone must've printed it out for you), NFL roster sizes are pretty much fixed-size, save a variable practice / scout crew and whomever you can stash on IR that isn't injured. Specialists are a necessary evil, but even so they're a premium. If one of the main goals of a NFL roster is the accumulation of talent in ways that can be well deployed, then there should be some value in players that can effectively cover two positions.
This is why I fail to understand how the Percy Position failed to catch on. Heck, even Percy Harvin isn't at the Percy Position for the Vikings. The Percy Position, in general, is a bit underutilized. Basically, this is a guy who can slot in well at WR or RB and get about 12-15 touches per game at both positions. Ideally, this should be about equal (targets are controllable, but only to a certain extent before it's completely implausible to force a pass into quadruple coverage - unless you're Drew Weatherford, of course), but the idea is to create a constant danger. Jet sweeps, direct snap play, screens, and the like work well when used frequently. Urban Meyer was a great proponent and user of this. Incidentally, two of the guys I consider to be among the best in the business - Chip Kelly and Gus Malzahn - don't really use this position. Instead, they opt for a heavy RB rotation to get effective carries, which is a luxury available at the college level. 7-RB rotations simply don't work in the NFL.
Enter Randall Cobb.
Cobb, of course, is a local product who played as close to the Percy Position as the SEC saw last season. He even saw some time in the Wildcats', er, Wildcat package. Cobb is as close to a legitimate run/catch threat as there is in this draft, and he's going to be drafted and used near-exclusively as a WR. While - again - I have issues with how the NFL pigeonholes skill sets, in this case this problem exists at the college level too.
Simply put, there aren't many truly transcendent players that need to get 10-15 touches per game. (Witness the ill-fated "G-Gun" package with Gerald Jones that was talked about for four months and used four times.) It's much easier to give lip service to the idea of a run/catch dual threat and run an incredibly obvious jet sweep once a game. That doesn't actually get you anything, though. Again, this is where read plays and limited use of the option comes into play. This is where a scaled, committed Wildcat makes sense.
Pat White was deployed in this package to decent effect with Miami, and give NFL offensive coordinators some credit for realizing it was a good thing. Then again, they thought the key was having a RB in the backfield; it was lip service by personnel formation. This was further compounded by every single NFL announcer calling any direct snap "Wildcat", which really didn't do much favors to raise the discourse level.
It's that lip service that's the issue; Harvin is dealing with it now in Minnesota. He's had carries, sure, but not the 5-9 per game he can handle, plus additional catches. If there's going to be a guy who's a legitimate dual threat, design and install a package designed to use him as such. Miami saw success with the Wildcat because they committed to it and spent time installing it and making sure that defensive coordinators had to respond to and respect it. If it's a gimmick, everyone will treat it as such.
Randall Cobb should succeed in the NFL. Whether he succeeds as much as he can is up to his offensive coordinator.