Hm, it sounds like they just got lucky, so i'm not sure what the lesson is here. A sample size of 1 is useless for learning what works and what not. I would be far more impressed if they would get similar results repeatably after each promoting action.