* Find a business that's getting good traction.
* Copy it mercilessly.
* Think of how you would improve it and implement the improvements.
* Profit!
Even if the final step might not be a monetary profit, you've accomplished something, and learned from the experience.
the pie is huge. taking small crumbs of marketshare as an individual is plenty.
might not be the model for your million dollar idea, but there's a subtle, guilty fun in beating people at their own game.
I somehow find it hard to recognize trends. I guess I am a hard core skeptical when it comes to noise v/s signal.
Setting up a Mechanical Turk for non-US markets might work. (As of today you can work on mturk as a foreigner, but you can not commision tasks.)
http://www.hit-builder.com/ offers a way around this problem for a $25 setup fee and 5% of use. I'm not familiar with mturk or this service but I suspect operating costs aren't that high. A barebones / follow on competitor could be a nice side project if you aren't concerned about amazon destroying your business model at some point in the future.