Relative TSR (rTSR) is increasingly used by market participants to judge and incentivize managerial performance. We evaluate the efficacy, reasons, and implications of firms' benchmarks in rTSR-based contracts. Although compensation consultants suggest that a primary objective of rTSR is to filter shocks unrelated to managerial performance, following the informativeness principle, we document that a significant subset of firms that choose index-based benchmarks do not adequately achieve this objective. Further, the index-benchmark selection is associated with governance-related frictions and is not driven by plausible alternative theories. Both structural calibration and reduced-form estimates reveal significant negative performance implications from suboptimal peer-selection.