Groups may outperform or underperform individuals, but no unified framework exists to explain these variations. I present a series of laboratory experiments focusing on sequential search—a general decision-making process where alternatives are sampled sequentially until an acceptable option is chosen. Key findings reveal that individuals consistently undersearch, sampling too few options and earning suboptimal payoffs. In contrast, groups outperform individuals when preferences are aligned (i.e., all members value alternatives similarly), due to both the mechanical aggregation of heterogeneous individuals and a "risky shift" in group decision-making. However, when preferences are misaligned (i.e., group members disagree on the value of alternatives), groups underperform due to collective search externalities, where group members' decisions to continue or stop the search may conflict. Additionally, voting rules significantly impact outcomes. These insights enhance our understanding of the relative performance of groups and individuals, even without private information or deliberation, and have implications for committee design.