How can Modeling Quality be measured / assessed?


Whilst a lot has been spoken about calculating ROI for Analytics by comparing pre and post implementation performance data points, very little content is available on measuring the Modeling quality.

The predictive power of an analytical model is a function of data quality used in the model. But with the same data different Modelers may use their skills differently to come up with models of varying prediction.

How can the Modeling Quality be measured / assessed then, given data quality as constant.