How do you design experiments and holdouts for enterprise programs given long feedback loops and smaller sample sizes?

With enterprise programs, I think you have to be realistic about what you can actually prove. You usually do not have the volume or short sales cycles needed for perfectl...