Assuming the potential performance of actively engaged people is roughly normally distributed (even if, as a group, they come from the upper end of the general

Assuming the potential performance of actively engaged people is roughly normally distributed (even if, as a group, they come from the upper end of the general population on that axis), then I think 50% is too conservative. Nobody can perform above their potential by definition, but lots and lots of people will fail to live up to their potential, making room for their below-average colleagues to outperform them with enough work.