Every startup should begin with the Business Model Canvas. They will refine this template as they go through the Lean Startup process and get their first Customer Development results. It is no longer optional for tech startups in any sector to use these concepts. They simply must apply them if they want to reduce their risk of failure. Steve Blank and Paul Graham have written extensively on how startups succeed. Reading their bodies of work saves founders a tremendous amount of time.
Accelerators now proliferate and some cater to specific verticals. Education startups have their choice of Kaplan EdTech Accelerator, Pearson Catalyst, Imagine K12, or DeVry's DV X Labs. Aspiring entrepreneurs should look at the track records of the startups coming out of those programs to see which one offers the best steroid power.
Workable education tech models need not be limited to MOOCs. Online education models are easily scalable but the barrier to entry is extremely low. The MOOC sector is already crowded and I believe the winners will be those that join established brands (i.e., the highest-quality private universities) to leverage accredited curricula. Other verticals may present higher entry barriers, and will thus offer durable competitive advantages to the first entrants.
I think physical hackerspace franchises like TechShop are extremely flexible by providing a la carte hardware capabilities. MOOCs are valuable because they make advanced academic knowledge universally available, but they cannot replicate physical laboratories. Let's say some MOOC partners with a hackerspace network to offer accredited courses in the physical sciences under an Ivy League brand. That market player would have incredible durability.
Identifying tools is easy. Finding an addressable market is harder. Education professionals who find their jobs frustrating are welcome to fight bureaucrats for the rest of their careers. The obvious shortcut is to escape from the ivory tower or bunker schoolhouse and launch a game-changing startup.