Because of the importance to spares management to the Flexible Sustainment program being proposed by Boeing to the US Air Force for the C-17 program (see Making the Case for Privatizing a Government Function), the task of creating an information and analysis system was put out for competitive procurement and the industry leader in logistics planning at the time was awarded the task. After more than 2 years of trying to implement a working system, however, Boeing fired the company, and Ventana agreed to build the system. Ventana did not accept the Boeing system specification, however, and built the system it thought would work. This system lacked several features in the Boeing specification, but it also had features not listed by Boeing. After 6 months, Ventana had a working prototype using 2 people (instead of 50) and applying a rapid prototyping approach to development. Over the following year, the system was improved and tested satisfactorily. The features not in the original Boeing specification turned out to be some of the most important. One was a portable implementation of the software that could be used in conjunction with the main software system. The other was a method of “cleaning” low quality data and keeping track of the cleaning.

The Air Force recognized that the ongoing task of finding and fixing data errors was the most challenging aspect of their logistics management activities. To meet the challenge, Ventana developed a data cleaning and logistics optimization system. The system combines a central database that facilitates changing, documenting, resolving differences and maintaining accountability for changes in the data, with a model of how parts status affects logistics system performance. By comparing model predictions to measured performance and automatically tracing the parts most responsible for the difference, the system identifies the data errors with highest impact on performance.

Evidence for the effectiveness of the system has emerged from real world tests. In a key exercise in December, 2000, twenty-five asset managers were relieved of normal duties for a week to clean the fleet data using conventional techniques. After this exercise, it took an analyst only five minutes with the Ventana system to sort through data on tens of thousands of parts and find twenty new errors, each of which was more significant to logistics performance than the errors identified conventionally. Most of the errors located the conventional way would have made little difference in calculating optimal numbers and locations of parts. On the other hand, without first correcting the twenty additional errors, any asset optimization would have been ludicrously inaccurate.

Perhaps surprising given the success, Ventana was not selected as the vendor to develop the spares management systems for other Boeing programs. Ventana was unwilling to make the schedule and cost promises that alternative vendors made. These promises were unrealistic, and they were not kept. The C-17 program refused to give up the Ventana system even after a corporate Boeing system became available.