The Customer

A major medical data processing corporation based in Chicago, providing brand-wise sales data for all the major pharmaceuticals companies in North America.

The Challenge

A very large amount of raw data was being collected from various sources, relating to customer drug purchases throughout the United States. It was required to develop a statistical analysis tool that would segregate these drug sales by arbitrarily defined geographical "regions", allow pharmaceutical companies to analyze customer spending trends in each region, and to set achievable sales targets for the sales force assigned to each region. It was also required to be able to set up an incentive system ("performance grid") that rewarded high performers, and to perform "what-if" payout analyses for different sales force performance scenarios.

The Solution

The raw data coming from a number of different sources served as the basis for the development of this software. Hence the first step was to ensure data integrity, and to consolidate the data into a single repository. An extensive study was done to study all possible source data formats, and to ensure that the final data in the repository was of the highest quality.

The second step consisted of studying the workflow of the sales managers responsible for setting the sales targets for their sales force, and for designing a user interface that they would find familiar and intuitive. Since the managers used Microsoft Excel extensively, support was also provided to be able to import data from existing Excel formats at any time. An initial user interface was created, and demonstrated to a cross section of the sales managers for their approval and sign-off before the actual software development commenced.

The statistical processing was accomplished by means of specialized statistical component libraries that provided very well-tested functions for most of the processing required. Use of these libraries not only speeded up the development, but also improved the quality of the software by abstracting the complexity of the numerical processing in the code.

Finally, in order to provide "what-if" analyses of possible sales scenarios, a Monte Carlo engine was developed in order to provide simulation capabilities. The entire software was developed and transitioned to the client in less than 6 months, allowing them to meet their commitments for the next sales cycle.