I've spent some time consulting for various projects, to be honest it's incredibly difficult to recommend even trying any more, to provide a service of reasonable quality you need seriously expensive equipment. On a small scale FTTP is probably more cost effective than FTTC, DSLAMs are expensive enough that GPON or similar offers better value (note: value, not initial cost) if you can avoid the having to dig up half the country issue. It requires expensive expertise to design, install and maintain (I do the design and oversee the install part and I know what I charge...).
You need massive take up figures, which means your offering must be substantially better or cheaper than the alternatives and even then you need to be really lucky. Seriously, you need every third or fourth property or similar levels in most of the models I see and that just does not happen.
You need to recoup the investment quickly - slowly doesn't work because mobile broadband will evolve into a threat within a few years unless you deploy *really* fast (aka really expensive) connections and other providers will see a big market in the area if you're successful and muscle in. You just cannot compete with BT / VM if they want a part of the market.
Recouping investment matters because you will need to re-invest within a few years to keep things working and there will be a big investment, you're talking serious band loans or venture capital funding in my experience. This isn't something you raise £50k for locally.
Support is a nightmare, because you provide internet connections you can't support them online, hence you need telephone support and it must be close to 24/7 to keep users happy. That's unfortunately expensive, many queries will be trivial and due to users but some will be your problem and hard to track down, you'll require a core of really talented people to find them.
Most of all, you basically need to contract a really good infrastructure consultancy with experience of dealing with this sort of thing. I run a company who do this kind of work and hence I know what we charge, which is influenced by what I can charge ISPs to do this sort of work. Basically, if we're not getting at least £600/day it's not worth getting out of bed.
Sorry that's incredibly negative I know but there are incredibly few places or situations where this sort of thing is viable today, partly because the experience required to do it is highly commercialized and even if you run the project as a not for profit then you'll still get a massive bill from your technical consultants. If you live on a small island or a remote community in scotland, there's some potential. If you just live in a home counties village BT can't see profit in upgrading then there isn't as a rule.
* Caveat - I'm aware there is a lucrative market in (local)government grants for these projects. I don't play those games because:
a) the grants are subject to how the government is feeling and in the current climate...
b) they usually come with sufficient terms to prevent you implementing a solution which is otherwise technically or financially optimal
EDIT: To be clear, I know the rules on self promotion and I'm not promoting my businesses consultancy services here, I'll happily answer basic questions here or by email (I prefer here if possible as it creates a record for others to refer to) but I won't get into too many specific details as I make a living from doing that and not from doing it for free.
