Memetics, Marketing, and Sociology can only get us so far.
I'm going to be very, very careful in this post; I don't know how to present this in a way that doesn't make me sound a little eccentric. I've been thinking a lot lately about memetics, the study of the transmission and propagation of ideas. These ideas have been compared alternately to genes and to biological viruses, spreading about hither and yon. There's been a bit of work done in the area of what ideas stick and what ideas don't, (consensus: various values are important, truth is only one of them) but there are just some huge, fat, gaping holes in the current state of the study of the propagation of ideas.
Gaping Hole #1: Knowing What We Can Know
First, and this is the scariest, we just have a very limited ability to view memes. We can see the spread of macro images on the internet, and identify that Ceiling Cat is probably sticking around for a while, whereas Rebecca Black's "Friday" is probably not. But we don't really know what to do with this information; the shape of the meme continues to elude us.
We will almost certainly never be able to predict what data will spread, and we will almost certainly never fully get a bead on what makes data "spreadable". We can get a general idea, perhaps, but it won't be an equation. Psychohistory is an idea in Asimov books, not a thing that will actually exist. (Oh, yeah, link. Sorry.)
This gaping hole is that memetic endeavors have largely been misguided because they've been strapped to a biological framework. Memes have no chromosomes, they cannot be put under a microscope, and they cannot be "sequenced".
Gaping Hole #2: Ideas and Behaviors
Second, there's not much in the literature by way of determining what, if any, is the difference between ideas that spread and the behaviors they leave behind. A piece of information that spreads might just annoy me by leaving "Never Gonna Give You Up" stuck in my head, or it might convince me to vote for Ralph Nader. There is a big difference between a meme and its behavioral payload.
Gaping Hole #3: The Myth of Measurement
Third, memetics tends toward the very abstract, and generally fails to measure anything at all, instead engaging in length thought experiments and exegeses on what theoretical construct is better for the task, in essence applying none of them empirically. Here's an article example, and it's ridiculously long.
The solution is to actually conduct experiments on actual cultural transmission. Richard Dawkins, the founder of memetics, once noted that memetics had not yet found its Crick and Watson; it hadn't even found its Mendel. I think he was wrong. Plenty of people have come before, studying memetics before it was called memetics. Most importantly, to my mind, is the sociolinguist William Labov.
Labov tackled issues of how language change spread, what factors made someone likely to adapt their dialect to another, and so forth. One of his earliest and most famous studies was one of employees in three different department stores in New York City. He found that one meme, the tendency to "correct" the New York City tendency to "drop" the r-sound in certain words (his test was "fourth floor"). The meme was more likely to spread along socioeconomic lines, meaning that those in stores with higher-priced goods tended to include r-sounds in "fourth floor".
I know this isn't really much, as far as studies go, but it was a start. Labov discovered a number of patterns of linguistic change, and scores of sociolinguists after him have followed suit. Sociolinguistics may be a little tame compared to full-on memetics, as language traits are hardly as world-changing as religious and political beliefs, but it is a sufficient, if terribly overlooked start. Not much different from Mendel's plants, if you think about it.