4 ms·

To this point, consider the failed startup Trada. (I worked there for a bit) They had a marketplace for ad campaigns with Google. Instead of running ad campaig

To this point, consider the failed startup Trada. (I worked there for a bit) They had a marketplace for ad campaigns with Google. Instead of running ad campaigns or using artificial intelligence, a pool of optimizers would work on a campaign and Trada worked as the middle man between optimizers and advertisers. The Trada logo was a jar of jelly beans: (http://wisdomofcrowds.blogspot.nl/2009/12/jelly-bean-experiment.html http://wisdomofcrowds.blogspot.nl/2009/12/jelly-bean-experim...) > A classic demonstration of group intelligence is the jelly-beans-in-the-jar experiment, in which invariably the group’s estimate is superior to the vast majority of the individual guesses. When finance professor Jack Treynor ran the experiment in his class with a jar that held 850 beans, the group estimate was 871. Only one of the fifty-six people in the class made a better guess. When advertising on Google competition leads to higher prices. For example the keywords "dui lawyer" cost a fortune because the rate of return is so high for each click. Theoretically a creative individual could come up with a set of untapped, low price keywords that could lead to an effective ad campaign at a lower price. In practice the company had major problems. Incompetent optimizers would chew through a whole budget, leading to advertiser churn, and a lot of intervention. Eventually everything becomes micro-managed which isn't a sustainable business model. (There's a reason agencies target big ad campaigns... its the only way to make it worth it because of the amount of resources you have to invest) So there's a single data point where the wisdom of the crowds failed.

kbenson 8y ago
I'm not sure I understand how incompetent optimizers leads to an example of the wisdom of the crowds failing. It sounds like the failing you describe is the result of individuals. Is there something I'm not getting?
- cdoxsey 8y ago
  The optimizers were the crowd in this business model. Some of them made good decisions, others made bad. But since the amount of money available was limited, those bad decisions brought down the performance of the campaign and ultimately the advertiser gave up. An expert would've made better decisions. I think the wisdom of the crowd can sometimes be effective, but there are a lot of ways it can go wrong. Some things I was thinking of: 1. the crowd wasn't big enough 2. the time frame was too short to be effective 3. the heuristic was too complex or poorly tuned to lead to a good outcome 4. creative endeavors with high rates of failure lack enough signal to properly optimize (a strategy might look terrible till it suddenly works) 5. misaligned incentives undermine the goal To illustrate the last case, in high school I had a teacher tell us that he would scale the test by the highest score. For example a 95 would give 100 and +5 to every other score in the class. With incentives like that, if everyone in the class answered no questions, everyone would get 100. (but then all it would take would be one individual to screw the whole class) Crowds are complex, difficult to understand and hard to predict.

kbenson 8y ago

I'm not sure I understand how incompetent optimizers leads to an example of the wisdom of the crowds failing. It sounds like the failing you describe is the result of individuals. Is there something I'm not getting?

cdoxsey 8y ago
The optimizers were the crowd in this business model. Some of them made good decisions, others made bad. But since the amount of money available was limited, those bad decisions brought down the performance of the campaign and ultimately the advertiser gave up. An expert would've made better decisions. I think the wisdom of the crowd can sometimes be effective, but there are a lot of ways it can go wrong. Some things I was thinking of: 1. the crowd wasn't big enough 2. the time frame was too short to be effective 3. the heuristic was too complex or poorly tuned to lead to a good outcome 4. creative endeavors with high rates of failure lack enough signal to properly optimize (a strategy might look terrible till it suddenly works) 5. misaligned incentives undermine the goal To illustrate the last case, in high school I had a teacher tell us that he would scale the test by the highest score. For example a 95 would give 100 and +5 to every other score in the class. With incentives like that, if everyone in the class answered no questions, everyone would get 100. (but then all it would take would be one individual to screw the whole class) Crowds are complex, difficult to understand and hard to predict.