When your coverage fails, this happens.
Learning a new scheme - offensive, defensive, whatever - is hard. It's not as hard if you're not starting from scratch; for example, it's not the hardest thing to teach a new route to a receiver if his position in the formation is changing, and it may not be the hardest thing to teach blocking schemes for the same play run out of multiple formations (power is power is power, and the blocking schemes at the point of attack are largely the same). Now, if you're doing something like going from a pass-first spread team to triple option? Get ready for a learning curve.
Now, you can learn a lot from reviewing film and studying playbooks, but some of the crucial keys that make a defense go can only be learned one of two places: the practice field and/or the game. So, when you're doing things like going from Justin Wilcox's conservative multiple-40 defense to Sal Sunseri's Saban-infused 3-4 with heavy doses of aggression and a whole new approach to pass defense, you need reps. Lots of reps. Enough reps that practice may not be quite enough.
And you need reps because pattern matching and the communication keys used are complex. And to even get to pattern matching, we need to go back in time a bit.
Man, Zone, and Evolution
So, broadly speaking, man coverage (I cover you) and zone coverage (I cover this space) are pretty easy to figure out and pretty easy to teach. However, they're also pretty easy to break down by sophisticated offenses; you can run routes that sap the DB's leverage (man) and throw to spots (zone), and so defenses needed to adapt.
Enter man-zone. This is a hybrid coverage type where you're ostensibly playing zone, but you're playing man coverage on anyone in your zone. So, for example, if you're a LB covering the left middle of the field, you'd cover guys like WRs from the left side on shallow crosses and meshes coming from the CB's zone. Now, you'd know they're coming your way thanks to communication between you and the CB. As coverage concepts go, this gives you the advantages of man and zone coverage, but there's an added cost: communication.
But wait, we're not done yet. Man-zone is still a reactive defense - that is, you don't know what the opponent is doing, only that he is doing something and you have a framework to stop them. A skilled offense can still take advantage of this by taking advantage of the inherent communication delays in the system, so there needs to be a better option.
Now, there's a bit of a little secret with setting up a passing game; there are a finite amount of routes offenses can and will run in the course of the game (and quite frankly, some combinations and routes are used more often by some offenses - think Mike Leach running the tar out of mesh), and so you can eventually learn that when X lines up here and Y lines up there, and X aligns this way and Y aligns that way, that X and Y are going to run a smash route, so you can now have your coverage align to take away that smash. That's the general idea behind pattern matching concepts - I know what the offense is going to do (or I at least know it's one of a few things), so my defense will be designed explicitly to limit those things. Can I still get burned? Sure, if it's not what I expect or I don't execute the defense well, but I should have more of an advantage here than I would otherwise.
With pattern matching, ideally you get the best of all worlds - you have a proactive defense that will actively work to take away what the offense wants to do, provided everything goes well. It's a nice solution, isn't it?
There are a bunch of details that go into developing a good pattern matching defense with sound principles, most of which would require a 17,000 word post (or you can just go here for some details on what Saban does), so for now I'm going to focus on the general things that come with developing and implementing a complex defense: language, repetition, and time. (You also need a guy who can act as the enforcer and key communicator. Minor detail.)
Nailing Down the Details
So, in order for this defense to work properly, you need to be able to communicate across the entire D. Basically, you need a call for most similar situations, calls to alert LBs and safeties about guys who may be coming into their area, calls to shift coverages, and you need to be able to a) recognize the offense, b) determine the solution, and c) communicate that solution, even if d) the offense is in hurry-up mode. Sounds like you need to learn the language first.
As far as what Sunseri's language is: I don't know. I imagine it's probably somewhat similar to Saban's terminology (see that link above for more than you ever cared to know), but Saban's terminology is extensive and probably requires flash cards and training to learn. We also don't know if Sunseri installed everything in one go or took a phased installation approach. Still, even if he installed most of the core components of the defense already, that's a lot of terminology to learn, let alone practice and execute.
And that's kind of the issue now; there's a big ol' practice time limit that will serve to limit Tennessee's learning curve. Once they know the language, it's simply applying the language in the right situation, but that takes lots of time to get right. You can get some of that in thanks to practice time, but you need game reps to really get it right.
As fans, it's one of those things that isn't obvious - unless it breaks down. We know then for sure, because this guy is just streaking down the sideline with nobody in sight oops.
So, if you're wondering why a cupcake like Georgia State ends up on the schedule, think of it like practice that counts. It'll be an opportunity for the secondary to get things right more before Florida comes to town and the real meat of the SEC schedule starts up.