An article in the Wall Street Journal (Jan 4,2012) describes Schwan Foods network of        6,000 home delivery drivers and their use of real time data analytics. Operasolutions, an analytic consulting company, mines purchase patterns to identify a genetic twin of a customer i.e., similar purchase patterns and  uses that data to recommend products. Similarly, when purchase within a category drops, discounts or recommendations from other categories can maintain spend patterns. The suggestions are sent in real time to driver handheld devices and have improved revenues between 3 and 4 %. Does the observed impact of real time analytics reasonable to expect in other contexts ? Should the goal be to maintain spends or move customers to higher margin items ? Should these recommendation services be charged for separately i.e., separate from the product costs  ?