In the K Means Cluster platform, the Cluster Comparison report gives fit statistics to compare the various models. The fit statistic is the Cubic Clustering Criterion (CCC). In general, larger CCC values indicate a better fit. However, fewer numbers of clusters can be more interpretable. The best fit is identified as the number of clusters with the highest CCC value, while favoring fewer numbers of clusters. The best fit is indicated with the designation Optimal CCC in a column called Best. For more specific guidelines on how to interpret the CCC, see SAS Institute Inc. (1983). Constant columns are not included in the CCC calculation.
Caution: If clusters are elongated or irregularly shaped, the Cubic Clustering Criterion should not be used as a clustering criterion.