How Stuff Spreads #1: Gangnam Style vs Harlem Shake
facegroup.comI think the article brings up a lot of great points and analysis, but misses the mark on one of the key foundational elements that makes these videos so different. Gangnam Style and Harlem Shake are different types of media. Gangnam Style is a 3:40 music video by an actual artist with a dance. Harlem Shake is a dance to a 30 second song clip. They both include music and dance, but the length and other factors make them very different and in some cases unworthy of comparison.
So from the beginning you're looking at two different lengths and types of content. Is 30 seconds more "shareable"? Probably. Does a pop song climbing the charts earn the music video more publicity? Probably. Can you then compare lifetimes of these two memes against each other and draw conclusions? No probably not. It's apples and oranges.
Maybe it's just a bad infographic, but I feel there are a lot of valid conclusions that could have been drawn from this type of analysis. But instead of that you get comparisons without explanation of why these memes are different and have different characteristics. I think they could have looked at two long-form music video memes or two short 30-second video clip memes, drawn better comparisons and put together a more complete and accurate analysis of these memes.
>video by an actual artist
PSY is a joke from the beginning, as is LMFAO et al. Harlem Shake was originally just a song which got somehow mixed up with a crazy dance and that went viral.
Pretty nice article about how "viral" at all the harlem shake was. http://qz.com/67991/you-didnt-make-the-harlem-shake-go-viral...
Yes! I remember reading this, it is notable that Facegroup failed to mention this, or to even spot that the "viral" nature was being pushed by particular groups.
Oh, "Each video was led by an individual or organization with massive reach – YouTube", yeah, real deep analysis. Much more interesting would be to break down into who posted what videos, who was hyping them on twitter, and so on.
Could the influencers and their actions described in the qz article be spotted by some form of algorithmic analysis? Now that would be interesting.
it is notable that Facegroup failed to mention this, or to even spot that the "viral" nature was being pushed by particular groups
The first rule of viral marketing club is that you do not talk about viral marketing club.
As some parts of the Internet would put it, Harlem Shake was a "forced meme".
This is conjecture at best, snake oil at worst (considering this is an agency that sells its research.) The data points are interesting, but the conclusions simply do not follow from the evidence. There's an implicit assumption that the actual 'value' of the videos is the same, that the content is ultimately fungible and what matters is how the content was originally shared and by what communities. However, to even the most casual observer, the 'memes' are two very different things.
The analysis claims that Gangnam Style had a "leader" and Harlem Shake was distributed. However, this is a kind of warping of the fact that Gangnam style was viral and satisfying in its own right, whereas much of Harlem Shake's value came from the parody videos and the fact it was a "thing people are doing."
There's no mention of the fact that Gangnam Style could and did make it to the radio in recognizable form. There's no mention of the length of the video, the season they were released, the "singability" of the content, the production value, actions of pr/production agencies, or countless other factors that could have a larger effect than the identified parameters.
And they failed to spot that Harlem Shake certainly did have one or two leaders - see the qz.com article linked by sweedy in a previous comment.
Wow, that missed the mark.
I always thought big media invented the harlem shake, because they failed to see Gangnam Style rising for several months.
Statistics would help us believe this article. Are they quoting 196% variance over the mean? What does it mean to be 4.5 times smaller? Even the basics are amiss here. There are a lot of conclusions from just two examples of highly complex dynamical systems. The authors poked around and thought a lot, and we know that unfortunately isn't enough to say something believable.
They explained what they meant by variation in point #3. We quantified this variation by first calculating the standard deviation of the daily sharing rate (i.e. how much sharing levels varied day by day), then dividing by the mean to give us the coefficient of variation.
Ah, I missed their explanation. The 196% and 338% are still bizarrely high figures, given that the variance of a Poisson distribution equals the mean. 338% variance, and you don't ask what kind of distribution you're measuring?
Mirror, as it's currently down for me http://www.facegroup.com.nyud.net/how-stuff-spreads-1-gangna...