Even from a Bayesian perspective, having a small sample that is too small to reach strong conclusions does not shift your posterior towards 0.