Viral marketing campaigns
| Gabriel |
Ad Age has a very interesting article on viral marketing campaigns today which graphs the views over time of various videos created by corporate advertisers or social movements. The most interesting thing about it is that only one of the campaigns shown in the graph (“Durex: Get It On”) shows anything like a classic s-curve. The rest of the campaigns can broadly be lumped into two categories.
The first type of video follows a pure external influence model (e.g. “Guitar Hero: Bike Hero”) and so they were only viral in aspiration but not in fact. These videos have an initial burst of views, then top out at about a half a million views and stop. What seems to be happening is that the companies buy a bunch of banner ads, “featured video” slots and that sort of thing and a lot of people watch but almost nobody ever sends the link to their friends.
The second type of videos follow Bass’s classic mixed-influence model. These videos also seem to start out with banner ads and such, but they actually do get people to forward the link to their friends and so they ultimately become extremely popular (millions of views).
I think there are interesting findings from this research but also an opening for further work. The findings are further support for the big seed model. Most of these campaigns did not start small like a true epidemic but had an initial marketing push. Furthermore, the research supports the hypothesis that the asymptote is correlated with the endogenous component of the hazard. The opening is to figure out of those campaigns that had an initial push, which actually went viral? Was it something about where the video was hosted (eg youtube vs it’s own site)? The length of the video? Whether it looked slick or was a simulacra of lo-fi “user content”?
Also see similar research on diffusion models for web videos by Riley Crane.