We document that investment committees of major VCs use a voting rule where
one partner ‘championing’ an early-stage investment is sufficient to invest. Their
stated reason for this rule is to ‘catch outliers’. The same VCs use a more conven tional ‘majority’ rule for later-stage investments. This evidence points to a model
in which voting partners get signals about different project dimensions and super star projects excel on some dimensions even if flawed on others. In this case, if the
distribution of project values is sufficiently heavy-tailed, as for early-stage invest ments, a champions rule is optimal, while a greater degree of consensus is optimal
otherwise.