>I think agreeing on outcomes is a much harder issue than you think that it might be.

This. People will pitch their idea going into the experiment then torture the resulting data until it suggests that their proposal was the most effective.