Ore variability remains one of the most persistent challenges in iron ore operations, influencing blending efficiency, plant stability, recovery performance, and final product specification. Variations in iron grade and impurity levels arise from geological heterogeneity, mining process, and stockpile handling practices. Traditional ore characterisation relies heavily on in process sampling and laboratory analysis, which, while typically unbiased, provides delayed feedback, limited temporal resolution and may have worse precision than required. As a result, short-term variability within ore streams is often averaged out, detected only after it has already affected downstream processing or missed entirely. This paper presents a methodology for managing ore variance through continuous ore characterisation using on-belt Prompt Gamma Neutron Activation Analysis (PGNAA). The approach enables high-frequency measurement of key elemental parameters, including Fe, SiO2, and Al2O3, directly on the conveyor belt. Real-time visibility of ore composition allows operations to detect and respond to grade variability as it occurs, supporting proactive process control and stockpile management decisions.