Jim Hines has some interesting things to say about chasing the market: at writers workshops he (and I) are often asked, by folks in search of success, "What's popular right now? What's the next Big New Thing? What are agents and editors looking for?"
I'd like to remind any aspiring writers reading this right now that the only reason for paying any attention to the current runaway success is that you should avoid writing anything like it. Here's why:
The production pipeline at a major publisher runs on a 12 month production cycle (with quarterly or thrice-annual batches of new material being pushed out to the marketing/sales force), but finding a slot for a new author can take longer—my first sale to Ace was followed by book slots at 12 month intervals but it took nearly two years after acquisition (2001) before it came out (2003).
Furthermore, turnaround time for those editors who read slush, that is, unsolicited submissions, is abysmally slow: even if you're agented, if you're a new author you probably had to go through your agent's slushpile. So, call it 12-24 months from first submission of your finished manuscript to getting a contract—if you're successful in selling that first book. (Most authors aren't.)
Then there's writing time. If you have a day job to hold down, you can't write as fast as if you're writing for a living. I used to take 3 years per novel when I had a day job. (I now aim for an average of 2 novels/year.)
Upshot: it takes about 5-6 years to write a book, run the gauntlet of agent and editorial processes, and then get it into a publisher's production pipeline. If you self-publish, you can shave 2-4 years off this odyssey. You also shave off a lot of essential quality control in the process, and lose access to marketing resources. Unless you write like a maniac and don't bother editing, you're still going to be behind the curve.
Furthermore, a publishing trend takes some years to become established—Harry Potter, Hunger Games, et al were not overnight successes: they took 2-3 years each to build, and the derivative follow-ups (from writers with existing contracts and slots to fill) took another 2-3 years to begin filtering out of the pipeline after seeing the success of the type specimen. Someone trying to break in for the first time is therefore going to be not 12 months behind the bleeding edge, but more like 5-10 years.
Even if you've got a book slot to fill and the flexibility to chase the latest trending hot topic, it's a fraught affair. I'm working like crazy to do an SF thriller trilogy that riffs off the NSA/Snowden revelations: I was in the right place at the right time with a noir near-future thriller under contract, and it wasn't a great leap to update my plans. Nevertheless, it'll only show up on bookshelves in mid-2015. Hopefully ubiquitous surveillance and paranoia about the global security regime will still be a hot topic two years after the Edward Snowden story broke: otherwise, I'm stuffed.
Upshot: novel-length fiction is not an appropriate vehicle for commenting on current affairs, and trying to chase annual trends in novel-length fiction is just plain silly.
So what should you write?
Firstly, don't chase the current market trends: by the time you get there you'll be behind the curve. Instead, try and go where there's demand but no supply. (For example: I noticed around 2005 that there was an eerie shortage of near-future SF—as if millennial anxiety intimidated everyone into tip-toeing away from that area. Hence "Halting State" and "Rule 34".)
Secondly, be flexible: have a number of different ideas ready to write. Ideas are easy, execution is the hard bit. If you're set on writing zombie romance and there's a lot of it on the shelves, you'll probably have trouble selling a new novel in that micro-genre because it was the hot new thing 2-3 years ago. But if zombie romance is just one of your strings, you can focus on something else.
The laser-sharp focal point of my writing is crime. And humour. And espionage. And Lovecraftian horror. And space opera. And time travel. And politics. And urban fantasy. With side-orders of LGBTQ interest and traditional romance and hard SF on top. This has worked for me because if something comes into fashion and then goes out again I've still got something else that can sell. (The New Space Opera was hot in the late 1990s. I sold "Singularity Sky" in 2001. If I'd stuck with that series, and written nothing but, my career would probably be in big trouble by now because the New Space Opera is old enough to vote and is no longer terribly sexy.)
Third and final piece of advice: never commit to writing something at novel length that you aren't at least halfway in love with. Because if you're phoning it in, your readers will spot it and throw rotten tomatoes at you. And because there's no doom for a creative artist that's as dismal as being chained to a treadmill and forced to play a tune they secretly hate for the rest of their working lives.