Avatar

Hi, I'm Aaron Richard. You might remember me as that guy from some website who goes by the name "Ralph The Magician."

I live and work in New York City. I design. I code. I like to create things. I like radio, tea, and sailboats.

This is one of my blogs. I use this to showcase some of the things I've worked on that I think others might find particularly interesting. I maintain two other blogs: Blogs by Magic & Rants by Magic.

You can learn more about me, follow me on Twitter, or send me an email.

The social advertising experiment

A few years ago I tried to start a company. The name of the startup was Anatomy Ads. You can see the remnants of a functional prototype at anatomyads.com, if you wish. The purpose of this company was to find a new way to monetize new media.

The idea was to create a truly social advertising network around the idea of “variable CPM.” You would pay ‘x’ dollars for an ad, and the number of ‘y’ impressions you would receive would vary based on how many other people were buying into that space, and at what price. If people could enter an ad unit, at whatever price they wanted, what would happen? Would you have hundreds of people offering small increments for hot properties, or a few big players trying to dominate the pool? How would people act if they didn’t know how many impressions they’d end up getting? Would you ever reach the equilibrium price? These were the kinds of questions I wanted to answer.

Unfortunately, I never really found satisfactory answers to the questions I sought. At least, not yet. This post is about how it worked, what went wrong, and how I still very much intend to find the answers to these questions.

To clarify, the system worked like this:

  1. A person who creates content (say, a blogger) would sign up and post the widget on as many (or few) of their sites as they liked. The more sites they posted the widget on, the higher the number of aggregate impressions they’d generate, and the greater price premium they’d be able to command.
  2. Anyone (e.g. brand, another blogger, your best friend) could come in and “sponsor” your content through the widget. They could pay whatever price they wanted. The minimum (for the sake of transaction costs) was $1. There was no maximum. A sponsorship was good for 30-days. All of the sponsorships within any given 30-day period were pooled together.
  3. All ad units were 125²px square buttons. The reason being that even a normal person could come up with a piece of 125²px creative relatively easily by using an avatar or simple logo.
  4. The widget itself displayed four squares at any given time (horizontal, vertical, or square). The algorithm on the backend would determine what ads to display based on how many people were in that ad pool, and how much each person had paid.
  5. We (Anatomy Ads, the name of my company that built this platform) did not take a cut of any sponsorships. 100% of sponsorships would go directly from the sponsor to the content creator. It was our goal to absorb all the transaction costs. We were going to monetize the business by: (1) selling remnant inventory;  and (2) allowing larger brands and advertisers the ability to run traditional ads from time to time in a space that would be much more engaging.

The system operated in near-real time and it was actually quite fun to play with. You could see how many people were participating in any given pool and adjust your sponsorship accordingly. We did get a few hundred people to try it out and start playing with it. We would fill the system with fake money and then spend it as if they were poker chips, trying to feel out at what point someone would leave a particular sponsorship pool and look for another one. And if you were the first person in a pool, you might spend $1 and expect to get 10,000 impressions at the time of sponsorship, only to find out that 10 other people entered that pool by the end of the day, and you’d only be getting about 1,000 impressions now.

I still feel that it’s a great idea, but our particular implementation had a few critical flaws:

  • It should be the obvious by now — the system was too complicated. It would have made a better thesis than a business.
  • We didn’t have an embeddable Flash widget (although we did start working on one). As such, our unit was limited to sites, social networks, and profiles that support JavaScript. (i.e. no Facebook, MySpace, etc.)
  • We deviated away from IAB standard units. Huge mistake.
  • We forced people to create a separate login with us. This made everything more click-heavy than it should have been to encourage the kind of behavior we wanted. We should have used more open APIs. (Although, to our credit, at the time things like Facebook Connect and Twitter OAuth didn’t even exist.)
  • We didn’t have the capital or clout to compete with the closest competitor—the now defunct—TipJoy.
  • I tried to bootstrap the company on personal credit, at the height of the worst credit crisis the world has ever seen.

I say we, but as The Dude would say, “The Royal ‘we’! You know, the editorial…” While there were a few other people involved in this project, I was the one responsible for its failure as a business.

But I’m not done with this. I have mothballed it down for now, but I will return to it at some point in the not to distant future. I believe that a system like this can work. It just needs to be simpler. A lot simpler. I also believe that this would make for an unprecedentedly powerful platform for non-profits and testimonials, especially if a future version of this system uses Facebook Connect and Twitter OAuth to actually generate the copy and creative that forms the social ads of the future.