We have found the opposite - teachers will pay for software if it is made well and makes their life easier. The price point obviously still needs to be set appropriately.
Life was very busy when I was teaching, anything that saves time on busy work is appreciated.
My experience and market research echos yours, just to get another tally mark here. Differenct schools allocate discretionary budgets differently, and there is some variance also between elementary, MS, and HS in the same district. A good time to reach for discretionary $ is near the end of the school year, after they've spent 7 months being frugal but don't want to "underspend," in part for fear that they won't get the full budget the next year if they don't spend it this year.
If you sell to the district then you've got an order of thousands of units likely to repeat for several years. If you sell to individual teachers then you're selling units at a time and probably having to resell those units (or face more competition) if you've got a good product.
The crap results were probably due to the soldering messing around with the crystal properties. Using a pressure-contacts approach instead seemed to do the trick:
I use it once every couple of weeks to tweak seating plans for my classes as I discover new relationships between the kids.
Desktop software because the SA algorithm placed too much load on the server, it started off as a Ruby web-app. Also desktop because most school networks have pretty draconian filtering in place.
Life was very busy when I was teaching, anything that saves time on busy work is appreciated.