Teradyne is a 45 year old corporation specialized in testing equipment for transistors and other electrical components of the semiconductor industry. The company was founded in 1960 by two MIT graduates with a vision to bring a line of reliable, fast testing equipment to the industry. As the complexity and volume of components manufactured by their customers increased, so did Teradyne’s investment in research and development. By 2004 Teradyne had five business units, including semiconductor test, assembly test, broadband test, connection systems, and diagnostic solutions. Semiconductor test remained the largest business operation for the company, producing just under two thirds of annual revenue in 2004. The company had six major engineering operations across the United States, with three of those facilities housing manufacturing operations. Teradyne had also dispersed across the globe with smaller engineering operations set up in Japan, China, and Germany.
The Semiconductor Industry
Semiconductors cover a very broad range of devices which can be classified by 2 main categories: memory, and system on chip. Each type of device performs a different task in an electronic system, with its own set of complex manipulations that it performs on the electrical signals that come as inputs. As semiconductors grew over the years to become smaller and more powerful, minor flaws in the production process could prevent an entire device from functioning correctly. This resulted in a high demand for testing equipment that could determine if a component was functional or not. This simple idea is quite a complex device, as devices often have a very wide range of operations. This is where Teradyne comes in.
Teradyne’s product performs the testing of components to ensure that their behavior lies within the specifications outlined by their manufactures, essentially telling them if a component is good or bad before it leaves the factory. By 2004 Teradyne had become one of the world’s leaders in semiconductor testing equipment with over 6000 employees worldwide. Their biggest competition came from Aligent, Advantest, and Credence, who held much smaller shares of the market.
Customers tended to be manufacturers like Intel, IBM, Hitachi, and Samsung who were actively creating new products and were willing to spend nearly 3 million dollars on a machine. Over the years, the production process of Teradyne’s customers had dramatically changed. Since production volumes were going up, and prices coming down, it was imperative that the testing process not slow down production operations. Reliability was of great concern as testing was often considered a bottleneck in the production process and any downtime would be extremely costly. Therefore, customer service was a major selling point as well.
As technology of semiconductors became more advanced, the need for more sophisticated testing equipment increased. So each new device often required purchasing a new specifically built piece of testing equipment. Therefore it was very common for customers to rely on Teradyne’s past experience with their products in developing new equipment. Furthermore, it was very common that once a company was chosen to develop testing equipment for a particular device, no other company would compete in that area. The rapid development of new semiconductors required equally rapidly development of testing machines, which proved to be a difficult task. It was becoming more important that testing companies come up with equipment that was no longer limited to testing a single device, but rather could be configured to test multiple devices. This would in short provide equipment that was considerably more difficult to develop, but would prove much more cost effective for the customer.
Teradyne’s background was strongly focused on engineering as many of the senior managers came from an engineering background. Projects were driven mainly by performance, focusing on technical competence. Engineers were motivated to dive into tasks, and encouraged to prove themselves. The company emphasized the individual. Many employees considered themselves at Teradyne for the long haul. Long hours were normal. In early 1990’s, Teradyne experienced a major change in leadership with the new CEO Alex D’Arbeloff. Alex grew interested in the risk of losing Teradyne’s competitive edge due to quality and reliability concerns.
Although D’Arbeloff believed the team was exceedingly competent, he believed there were major operating problems regarding reporting and performance measurements. To minimize these problems, D’Arbeloff embraced total quality management (TQM). After 5 years of intensive effort, TQM principles were embedded into most aspects of work at the company and resulted in noticeable improvement in manufacturing quality and customer service. However, the engineering organization was resisting TQM because of the resulting late and over budget projects.
In 2001, Teradyne senior management decided to make a fundamental change in the strategy and aimed to create a single flexible tester platform. The project named as Jaguar to represent how critical it is to the company’s competitive strategy. It is noted that, “from the outset, it was recognized that the project had to execute flawlessly.” So it is expected that the Jaguar project will be, unlike the Teradyne’s traditional projects, more focusing on execution strategy including increased emphasis on up-front planning and design, reorganization of project team structure, the implementation of project management prospective, and formalizing project management tools to streamline the product development activities.
It was decided that the engineering teams in Boston, Agoura Hills, San Jose, Minneapolis and Portland to collaborate and work on this project. Each site was responsible for a core subsystem. However, each of these teams had already their own approaches and legacy tools they previously used. The teams were led by the O’Brien, a 24 years veteran of Teradyne’s engineering organization. A core team was formed by the leaders of each group. This core team used to meet on weekly basis to discuss progress, and to make critical technical and organizational decisions. The top management recognized the project as pivotal and strategic to Teradyne’s future success.
It was extremely important to have the product out in the market by mid-2004. That’s why the project definition and scope were one of the critical decisions. More time was spent in the early stages of the development process, mainly concept development and product planning. The objective was mainly to identify what to do and what not to do to hit the market window. Teradyne was committed to the project and provided all means of support and resources to ensure that the project gets delivered on time. Whenever there were delays, senior management were able to provide additional resources.
To get the needed funding, detailed analysis of the product requirements, specifications and architecture were done and presented to senior management highlighting critical subsystems, target performance and project execution plans. This was a challenge for the software group to provide detailed specifications and make major architectural decisions at this early stages in the project taking into consideration lots of uncertainties associated with the nature of software development. During execution of the project, the core team utilized the following project management tools:
Work breakdown structure
Critical path analysis
Earned value analysis
However due to the different teams’ backgrounds and level of experience using these tools, the core team spent more time debating the tools and how the info is presented in the tools and how much this reflected reality. It was obvious in the core team meetings that the teams responsible for the hardware subsystems were delivering on time keeping all major activities on track. On the other side, the software groups were running at only 50% earned value each month completing only half of the tasks originally planned. However the groups were on denial saying that they would catch up, but it never happened. In Sept 2003, the project targets were further stretched when Alphatech, one of the biggest customers, was approached by one of Teradyne’s competitors.
The project has to be delivered to Alphatech by end of March 2004 rather than end of June. In addition, Alphatech introduced changes in the requirements to match their specific needs. Teradyne had no option but to comply with the new target because the loss of AlphaTech account would mean a loss of more customers who might follow the trend. As the deadline approached, with the increased pressure on the software groups, more features were coming out the door before they were actually ready. However, after the delivery, the software groups spent more time fixing bugs to get an acceptably operational software. Things were now getting worse, there was a huge burn-down but there was no development progress.
Lack of Experience
The software group lacked the needed experience on the IG-XL operating system, as most of the system experts were involved in other project development (FLEX), Jaguar’s team lacked the enough experience that would allow them to develop the new system upon this operating system, as most of the developers had never worked with IG-XL before.
Because of the excessive load that was put on the development team, the team members were burnout, the stress levels were enormous, which lead to developing a buggy software for AlphaTech. To fix the software issues, the team took more than six months in the bug fixing stage, the team stopped developing new features and they just shifted to firefighting mode, and the development process was ignored in this stage.
Although software team output was running at approximately 50% earned value per month, the management didn’t act to fix this issue, and it is known, the more you postpone solving the issues, the more time it would take to be solved. In addition, the managers were more skeptic around the metric and didn’t pay attention to the team delays
Escalation of Commitment
When senior management and the core team realized the large delay in the software subsystems, they added more resources to the project several times without even analyzing reasons why the groups are behind schedule. In addition, there were no measure for the impact of the extra resources on the project schedule and how it will bring that forward
Project Management tools usage
Although the project management tools provided the management with enough useful data, they management didn’t respond to it. In addition, the software team kept playing with the numbers to adjust the project plan, but the things done on the tool were not practical enough to be implemented successfully in the real life. Others became more keen to make the metrics look good, rather than doing what’s more important to the project.
What should have been done?
Learning curve should have been estimated properly, as sometimes the software development teams tend to underestimate the learning curve of the new tools (which happened at Teradyne), and expect that the team would work efficiently from the first day, which is a managerial common error. Teradyne could have assigned some of the system experts within the Jaguar team, as this would help the team overcome the experience issue Extensive training should have been delivered to Jaguar’s team Teradyne should have recruited more resources from the beginning of the project to the software development team to avoid burning out its resources and be able to deliver the project on time; this could have been reached by better estimation of the tasks, and making sure that the team has the needed knowledge to achieve its goals
The management should have interfered as early as possible to minimize the effect of the early delays, they could have got Consultancy Company that could help them, and they should have analyzed the root cause of the delay and try to fix it. Although we agree that the project metrics are important, they shouldn’t be the goal, they are just indicators of the project status, people should work to get the job done, not only trying to adjust the metrics The metrics should not be the only way to evaluate team’s performance, other competences should be taken into consideration such as team work, positivity and being accountable. In addition, the project managers should thoroughly analyze the data extracted from the tools periodically, to be able to take quick actions in case of any deviations that could happen.
Way to move forward
Now, having the product delivered to customers, the following actions need to be taken: Provide training for key software developers to maintain and enhance the product Recruit expert software engineers to lead the software development process Document the current project status, the implementation details and all other activities that still need to be done in terms of bugs, new features or enhancements.
Unify the project management tools and provide the needed training on how to use them, and the information provided to and extracted from these tools Prioritize the list of bugs to identify the major and critical ones that need to be fixed immediately Prioritize the list of features to be completed and identify the ones offering a minimum viable solution Since the product is now in the market, the project team needs to plan a new release with the scope defined as the minimum viable list of features and the major and critical bugs.