Intra-class variance measures the variability of data points within the same class, reflecting how spread out or close the data points are to the class mean. Minimizing intra-class variance is crucial for enhancing the effectiveness of classification algorithms by ensuring that data points within a class are tightly clustered together.