Video blog that explains the rationale behind a poorly studied treatment (eg binaural beats, neurofeedback, RISUG), and how it could be tested. Then the ad revenue from the videos, plus crowdfunding, is used to actually perform the described test and analysis, and the results are blogged out. Idea credit to Rachel Lipson, necessity credit to NSERC.
Binaural beats as a treatment for insomnia may be a good place to start. There is an open source Android app that I could edit to administer a placebo. The placebo would be a real binaural that fades into a monaural after a short time. It could use some ID number like the SIM card to determine who gets a placebo, and could record how much it was used. A validated survey instrument for insomnia could be attached to the app, and voila! I have my data recording done.
The only costs would be licensing for the validated instrument and maybe hiring an app developer. If it was a success, it could spring into more expensive treatments like neurofeedback.