I think that Developers(D) don't want to work with a nebulous perspective of a payment. On the other hand, users don't want to risk prepaying without any guarantees of receiving code. I think that it would be better to distinguish users as Clients(C) who would only pay for code and Investors(I) who will pay to Developer directly but will get compensated for their risk after successful development.
Here is my scheme of crowdfunding:
Phase 1: Predevelopment.
Creating the feature.
Clients send money to a fund for ending (FE) (can be refunded).
Developers create offers of the required feature.
Investors send money to the starting fund of a developer (each has one's own) with request of wanted money after the development is a success (can be refunded).
Phase 2: Development.
The Developer chooses Investors if the fund for completing the project is greater then the sum of money requested by chosen Investors (difference will be paid after successful development) and declares start of the development.
Other starting funds are refunded to the Investors.
The Developer gets money from chosen Investors.
The Fund for ending is frozen for refund.
New Investors can send requests to developer. If the developer accepts them and the fund for ending has enough money, then the funded money is sent to the developer and Investors wait for development to complete.
If development terminates without achieving project's goals, the fund for ending is refunded to clients.
Phase 3: Release.
After finishing the code and fixing bugs in the feature releases.
Fund for ending is paid to investors and the developer.