Management and the Technology Professional – B302

Case study answer 2

If I were the supervising engineer responsible for step six in the manufacturing process, I would have recognised that missing the test could cause significant delays and expense when the product goes to market. The identified critical path is to fabricate transistors, perform quality tests, sort components, perform final tests, configure for market and finally to do product marking and localisation. There are currently two options which are available to me and the project. The first option is to delay the schedule, and conduct the test, although this would inevitably push out the final deadline. This option may also cause subsequent tasks to restart if started once step six is completed. Rerunning the test is a four day automatic process so all tests will have to be repeated on that particular batch. The time to configure the software to run just one test will prove more complex than to just run the entire test package. Alternatively, the final test could be ignored; however this would potentially cause the problems which the test usually identifies to remain in the product. These problems could then be found later, possibly by the customer which may harm the company’s reputation. Acting reactively is generally more costly.

Personal experience tells me that before a product is launched, it is crucial that the manufacturing testing is completed. At a previous company, I worked on testing printer drivers. The initial test verified all functionality as working. A new version was released which we skipped the regression testing to save time, we assumed that the features worked as they did in the previous version. The drivers were then released to a larger test audience, and problems were detected. The assumption resulted in the project losing more time as the missed regression test had to be repeated. Also all the testing completed by the larger test audience had to be restarted. My knowledge tells me that at Intel there has never been a case in which we have not completed the testing. Shipping the product without the final test completed is new territory and one which we should avoid. It is important to look at past test results to see what has been identified and the severity of the problems. The test normally identifies few if any critical, major and ordinary bugs and therefore it could be considered not essential. When considering the full project and that this iteration of testing is only for a single batch then the product should be progressed. Four days testing is 25% of the entire path which would cause significant delay. In the ideal world the test should be completed at this stage, as the earlier problems are identified the less money is required to correct the mistakes. By completing the test we know that we will have equalled the quality performance of previously shipped products.

All things considered as the supervising manufacturing engineer I would prefer the test is completed because my job would be more at risk if a product is shipped to customers with faults than if the schedule delayed. At Intel we always complete our testing phase and with this product there should be no exception. An action plan should be put in place to complete the test, overtime and extra heads from other departments should be considered to ensure minimal time is lost and the product ships as close to launch as possible. The extra effort now ensures the customer will experience no different in product quality to any other Intel product. Combating problems earlier is cheaper than fixing them once they are in the field.