The point would be experimentation. You do a small one, measure the results, and if it aligns with theoretical predictions, you scale up. It transforms it from high risk and high reward, into low risk high reward.