When Advice Can Harm

Abstract

Strategic decisions can be hard, and when facing hard problems, decision-makers commonly seek others’ advice. But finding a good advisor is tricky: People are biased judges of skill, even more so in the wicked environment that typifies strategic decisions, where data can be incomplete, outcomes ambiguous, and causality foggy. Relying on empirical data, we model when and how person-to-person sharing can cause bad advice (or ideas) to spread, even without conformity pressures or malintent. We begin by reviewing how people systematically misjudge relative skill: misestimating how their own skill compares to others’. We show that if skill misjudgment is assumed away, advice taking is always beneficial. But under realistic assumptions, advice sharing benefits the low-skilled while harms highly skilled decision-makers, degrading collective skill and potentially eliminating top performers. We test an organizational design solution: When decision-makers are clustered, paying more attention to close associates, the range of practices increases. There lies a trade-off: Encouraging advice exchange homogenizes skill whereas discouraging it benefits diversity. For scholars of organizations and knowledge, we offer a novel theoretical mechanism, rooted in extensive evidence. It is distinct from prior concerns about normative conformity, and thus may be resistant to the proposed remedies. For practitioners, we sound a warning against the risks of popular practices that are meant to increase interaction and advice exchange in firms, among entrepreneurs, and in crowds.