Data-Driven Football Performance Measurement: A Measure-First Systematic Review of Action-Centric, Individual, Team, and Financial Analytics (2018–2024)

Football analytics has produced many performance measures, including action-value functions, plus-minus ratings, tracking-based models, and machine-learning models of market value. These measures estimate different aspects of performance and are validated against different reference points. This systematic review consolidates data-driven performance measures for elite men's football published from 1 January 2018 to 31 December 2024. Following PRISMA 2020 guidance, we searched Scopus, ScienceDirect, and SAGE Journals, screened records in duplicate, and synthesized 74 studies. Measures were organized into the Action-Centric, Individual, Team, and Financial decision layers. Performance measurement shifted from event and outcome tallies toward context-adjusted models using spatial coordinates, possession context, tracking data, contract information, and market signals. Median factor breadth increased unevenly from 13.0 factors per study in 2018 to 22.5 in 2024. Additional variables were useful mainly when they added decision-relevant context or formed stable composites. Action-value measures were most useful for valuing discrete events. Bottom-up ratings improved player attribution in event-rich settings. Rolling expected goal difference was the most transferable team-level signal. Financial models performed best when performance indicators were combined with contract and demand-side variables. Measure choice should follow the decision context, validation target, available data, and explanatory needs in practice and research applications.