Marketing is where the predictive models usually generate the highest ROI, so it's where it's the more important to have a good tool. Also, in marketing, you usually end-up with datasets with several million rows and several thousand variables (our typical solution in telecom uses 2500 variables) and then the free solutions like R do not work anymore. ...and we are also using TIMi in many other domains: recommendation system, medical diagnostics, text mining, preventive maintenance, soft-sensors, etc.