Abstract: The standard simple sequential herding model is altered to allow a firm with a new product to have it reviewed publicly before launch. Reviewers are either inherently pessimistic, optimistic or unbiased. We find the counter-intuitive result that a firm with a good product will prefer a pessimistic reviewer. Although firms with a bad product prefer unbiased reviewers, signalling considerations will force them to copy the choice of the good product firm in order to avoid revealing product type. This asymmetric impact provides a strong explanation for the stylized fact that reviewers are often viewed as being very critical.