<p>Global university rankings play a central role in shaping institutional strategies, funding priorities, and academic reputation. Among their core indicators, field-weighted citation impact (FWCI) is widely used to represent research quality. Despite its prominence, the underlying mechanisms of FWCI and its influence on ranking outcomes remain insufficiently examined at the institutional level. This study investigates internal logic, empirical behavior, and possible inflation (disproportionally high values) reasons of FWCI, which can substantially affect institutional ranking outcomes. For that a dataset of 230,357 publications from ten highly ranked universities over the 2021–2025 period was analyzed. It was revealed that beyond the expected sensitivity of FWCI, such as citation baselines and publication maturity, other factors like subject classification systems and document types, exert a substantial influence on its values. Specifically, conference papers and chapters in some fields can have up to 5 times higher average FWCI-to-citation than articles, with maximum ratios above 2.2 and 2.8 respectively for the most mature years. Besides, for some fields and publication types, open-access outputs can exhibit up to two times higher FWCI values than non-open-access ones. The findings were confirmed by a validation conducted on a broader set of 100 institutions from the top 1,000 of the THE Rankings 2025. Conference papers with books and chapters reaching up to one third of total output can create conditions where the amplification translates into unbalanced institutional FWCI performance. To address this, a simple adjustment is proposed to reduce FWCI inflation and improve interpretability, reducing conference paper and chapter FWCI by around 50%. The study contributes to the ongoing discussion on robustness, fairness, and transparency of citation-based indicators and provides methodological guidance for refining FWCI-based institutional assessments.</p>