A guy at a conference was telling me how he really wanted to build a music recommendation service, but had been trying for a year to raise the $2 million dollars he said it'd take to build it.
My suggestion for him: Don't wait for funding. START NOW. Like this:
- get a dedicated phone number like Google Voice
- tell friends to call you at that number for music recommendations
- they call you, tell you what they like, and you recommend something they might like
- write down in a spreadsheet what they requested and what you recommended
- afterwards, ask how happy they were with your recommendation. write that down, too.
- eventually put this spreadsheet into a database
- eventually put this database on a website - letting people browse past satisfied recommendations
- keep improving your ability to recommend (by asking experts, learning more)
- eventually write a program to have the computer recommend without you
By getting that initial lo-fi hands-on experience, you'll have a deeper understanding of what people really want from music recommendations.
Then you can build your service incrementally based on real user communication, instead of hiding in a lab for a year programming in isolation based on a year-old hunch.
If you say you want to do something, DO IT! Never blame outside forces stopping you. Work around obstacles to start immediately.
(P.S. He didn't like my advice. That was a few months ago. He's probably still looking for funding.)