Jackson Miller

There is a great article in the current issue of Wired about using human intelligence to collect data that computers have been having a hard time collecting. This is not the same mundane crowd-sourcing shit as Mechanical Turk though. It is more like SETI@home in that it is kinda using “spare cycles” for the processing power. Make sense? (probably not yet)

Luis Von Ahn, the guy who created CAPTCHAs*, is now doing research on the data. One example is a “Hot or not” like game with pictures of various places and things. While the goal of the game is for multiple people to agree on the aesthetics of images, the goal of the research is to collect information about beauty for more work later. Currently there would be no way for a computer to determine which picture is prettier, but after a critical mass of data is collected computers may be able to calculate the pattern of aesthetics.

Currently, humans beat computers at determining beauty. I think another area that humans (so far) beat computers is in determining relevance. While determining beauty is “neat”, determining relevance is making people shit-tons of money. Amazon exploded onto the scene by making recommendations based on what other people who bought the item you are currently looking at also bought. This is very similar to the research that Von Ahn is doing in that it is observing the decisions of humans to better predict how humans will decide. Most product recommendation engines have taken a similar approach.

I am real big on the idea of software that collects data through passive observation and then uses that data to provide suggestions or new information to the users. The trick is to reduce the pain of humans entering the information. The problem with crowd-sourcing like Mechanical Turk is that it doesn’t reduce the pain and monotony of data entry, it just tries to reward the pain with money so small you have to question the cognitive ability of anyone who is motivated by it. The brilliance of the work of Von Ahn is that it eliminates the pain and even makes it fun. Talk about aligning motivations! Which brings me to magazines.

Huh? Magazines?!? Yep.

When I worked at magazines.com we spent a few months where we were constantly talking about blogging. “Blogs” were exploding onto the scene and as Newsweek started doing articles on blogs it became apparent that we needed a “blogging strategy”. I had an idea that was not adopted, but I am confident it had game-changing potential.

People who write typically like to read. My thought was that bloggers probably liked to read magazines. People who read blogs obviously like to read, so they probably read magazines as well. Since most people who write blogs about the things they are interested in, the people who read those blogs are probably at least marginally interested in the same thing. What I wanted to do was to capitalize on these aligned interests and make it drop-dead simple to turn bloggers into affiliates. The way to make it drop-dead simple was to allow bloggers to select the magazines they read and are interested in and we would do the rest. If I like Wired, chances are pretty good that readers of my blog like Wired. There is also the added benefit repetitive advertising for regular readers; eventually they had to buy the magazine.

At one point we put together a business plan for statzen that included self-selecting magazine affiliates as an additional revenue stream. I half-assed pitched it to magazines.com, but they were not interested. The profit numbers on that business-plan are both obscene and completely feasible. At some point I may end up funding statzen with a similar model**. Which brings me to Facebook.

Wha? Facebook?!? Yep.

The Magazine Rack Facebook App that Rex talked about is damn close to the app I wanted to develop for magazines.com. You know the main difference? I was planning to share the revenue with the bloggers (or Facebook users now that we are in 2007). Magazine Rack is spreading on the sheer interest of letting others know what magazines you are in to. The sad thing is that the developers of Magazine Rack are probably leaving lots of money on the table by using Amazon instead of one of the dedicated magazine sites that pay higher affiliate commissions (30-40% of every subscription). The sadder thing is that I coulda-shoulda-woulda developed this 2-3 years ago.

As the web becomes more social, there will continue to be new opportunities to use a little ‘work’ from humans to create a powerful relevance engine for e-commerce. Sure, you could write a recommendation engine for Netflix and have a chance of making a million dollars or you can find a interesting and relevant new ways to market affiliate products. The social-web will provide much greater visibility to the long tail and there is plenty of room for marketing innovation to capitalize on it.

* CAPTCHAs are those annoying image verification scripts when posting comments on blogs or signing up for an account

** Any help finding an angel investor would be greatly appreciated

Note: For a while I made hundreds of dollars a month with a magazine affiliate on jaxn.org. Unfortunately the proof came just after I left magazines.com as an employee.