For my (post) masters project on machine learning and big data infrastructure I thought it would be fun to acquire my own data set. Last semester I traded available services and architected a scalable (big data) Internet facing survey infrastructure using a combination of Amazon Web Services and Python. I had my friends send me (sometimes silly) survey questions and put them on my site. After an initial flury of "friends of friends" taking the survey, my traffic died down to zero.
I then decided to pull the trigger on web advertising and started by Google searching "how do I advertise on the Internet."
I signed up for four of the most attractive options. I won't bore you with the also-rans (although I should mention one Advertising service turned down my cold hard cash because my site "does not meet [their] partner ad networks' strict policies regarding site content, structure and navigability") since, as far as I can tell, Facebook's advertising service delivers the platonic ideal of perfection. I will try really, really hard not to use cliches such as "knocks my socks off," "blows me away" or "combines to form Predaking and destroy the Autobots" when I communicate how well this service works.
I'm sure if I (devote my nights and weekends and) master the art Search Engine Optimization, A/B testing and AdWords I would find fault with Facebook's service, but that is my point. Facebook gave me, someone who has never advertised on the Web before, a Fischer-Price interface to quickly deploy a killer campaign with no thought or skill. One of the other services I tried, in contrast, required me to install a "pixel" to my site, which is simple enough for me (once I figured out how to make it play nice with my Bootstrapped Jinja2 template), but I could never see Al (from Al's auto-parts) pull that one off. With Facebook's service, I could see a Caveman mashing his keyboard with a club deploy an effective campaign. I think that is why I like Facebook's service so much. Any "Mom and Pop" could figure it out, and Facebook is smart enough to keep things moving to their sale. Click, Click, Click and close. Genius. The competitors put up huge roadblocks that required me to Stack-Exchange some questions wheras Facebook takes the attitude of "Don't know what this means? Don't worry about it, you can tune that knob later, let's move on..."
To summarize the experience, you simply click "Create Ad," pick some demographics (they had one for "people who take personality tests"), create a "Call to Action" (A slogan to get people to click) and then upload a JPEG. I didn't have any graphic art so I Googled "Public Domain Cartoon Brain," uploaded that, and I was done. The results? 86+k Imprints, 7k clicks to my site and 4.75k of those taking the survey! All for just \$427. I will let this run for a few more days, just to see what I can get.
A quick note. Of the 4.75k test takers I had two skeptics that think I'm a front for some political entity or some other nefarious institution. I assure you that I am putting together this data set for my (post) masters college project. Also, as I said before if by some miracle I make any money off this, I will donate it to charity! For the time being, with the infrastructure and advertising I'm a couple grand in the hole and forsee a couple grand more. I am paying for this out of my own pocket. Where do I get the money, you ask? Easy... by selling my childhood on ebay!
UPDATE: I posted this blog entry at 2:04 AM yesterday. Since I posted this blog, the market opened and Facebook stock went up 15.5%. I take full credit for this jump: