I was in a similar situation recently working with minimally funded organizations. In the end, after spending about two years on it, I wasn't able to make it work. The groups' needs were just different enough that they required consulting engagements but couldn't really afford them. I've found it disappointingly hard to serve small businesses and NGOs given that they often do interesting and important work. I hope you can make it happen. Some thoughts:
0) There's a good chance your main obstacles aren't related to server optimization or finding the right mix of features to roll into a product. You probably need to clarify your assumptions and make sure there's really a business to be had.
1) Consider honestly whether it's realistic to serve organizations that can't pay $40 a month for software they need. I'm not saying it's impossible or undesirable, but it's a critical question to answer for yourself. You need to eat, and choosing to work with these organizations means you'll have to have to subsidize your time with other, higher paying work, offer a level of service that won't satisfy you or your customers, or both.
It's possible there's a large enough number of customers that you could support a business on < $40/month, but it's unlikely for non-profits, especially given your experience with their heterogeneous needs.
2) In light of their reluctance to spend, how sure are you that you've identified a critical need for these organizations? This isn't necessarily an indictment of your product. In my experience many small NGOs are still figuring out their internal processes, and what's important to them and their funders. Codifying nascent processes in software is hard - bordering on impossible - and sales cycles can be untenably long. On the plus side it means they're often willing to be flexible and work with you as you figure out what the software should do.
3) If you do move forward (and of course you probably will), be creative about funding. Working with non-profit arts organizations gives you a great story to tell. If submission management and evaluation really is critical, can you work with your customers to incorporate a software budget into their funding proposals? Find out who the funders are. What are their priorities? Can you appeal directly to funders and show them how your software can help achieve and monitor their objectives, and communicate their impact more effectively?
I'd love to hear what you end up doing. Feel free to reach out (my username at gmail) if I can help.