I think the theoretically optimal way to distribute credit would be to compute Shapley values for each contributor (including the the founder) and distribute the funding accordingly. But this is of course usually not possible (because of limited information).