Perhaps an example of how "small" a trend can be could excite you enough to look out for things that are "bite sized" enough to try.
I'm pretty deep in the Ruby world. A trend that has emerged (and is almost becoming a cliché) over the past 18 months or so is selling Ruby related screencasts. Peepcode started selling screencasts about 3 years ago but no-one followed on in a serious fashion for quite a while (until Peepcode "proved" it could work, I say).
Now, there are at least 7 - 8 unique vendors of Ruby related screencasts I can think of. None are Peepcode scale at all, but I've seen sales figures and know that some of them, at least, have done reasonably well. Even the people behind the more humble attempts are happy, see: http://codeulate.com/2010/03/how-to-sell-a-hundred-screencas...
Now, "Ruby screencasts" is a crazily tiny niche. Not only is "programming" a tiny niche in the world of business, "Ruby programming" is a MINISCULE niche in the world of programming.. and Ruby programming screencasts is ULTRA CRAZY MOLECULAR LEVEL TINY!
If people have picked up on a trend in such a miniscule market and are making money with it, that's says, to me, that there are millions of teeny tiny niches in which you can start to perform very simple "tests" that make money and which, ultimately, could lead you to that million dollar concept or market.
Picking up on these teeny, tiny "trends" can be tricky, but is a lot easier in fields you're already absorbed in. If I had to pick out some trends just from reading Hacker News recently, these seem like a few areas someone could make some bucks in now/soonish: node.js consultancy/screencasts/books, Posterous theming, Flash-to-HTML5 tools/consultancy/screencasts/books, non-college oriented compsci education.