This is the opposite of the truth for small businesses: you can sell to any teacher, all you have to do is convince her to get out her credit card. To sell it to a district, you're going to have to go through a sales cycle measured in months to years. You will have to collect rubber stamps from upwards of four different decisionmakers, three of whom have no interest in using your software, at least one of whom has incentives to kill the purchase.

Selling software to educational institutions is largely a matter of playing Enterprise Sales. I don't suggest you make that choice lightly.

I agree with the above - in its current form, I think you'd attract early adopters who, even if they love your product, might have trouble convincing others to sign up and fork out for it. With an institutional model, said early adopters (presumably those in charge of educational technology spending) can try and push it out to teachers in their school.

Along these lines, perhaps consider some way to allow those who use and love your product to bring others in with minimal friction - let them try out the product in an administrative and student role, without having to sign up and learn on their own. Most teachers I know won't try something out until they really see the value in it (i.e. students engaged, fulfills some kind of need).

We're working on adding institutional accounts. There are some complexities involved, mainly involving legal issues and various whitepapers that need to be written, but I wholeheartedly agree with your idea: early adopters are a great way to get institutions to adopt a product.

As for your second suggestion, do you think a way for current users to send potential users an invitation to try a demo account (a sandbox) would do the job?