Rockwell Automation's authorization procedure systematically guides a product through a four-stage development process

Rockwell Automation in Mayfield Heights, Ohio, is now one of the three product divisions that remain within Rockwell International. A leading manufacturer of industrial automation-control products, Rockwell Automation was officially created in 1992, when the parent firm renamed its Allen-Bradley Co. subsidiary. Allen-Bradley, which Rockwell had purchased in 1985, is now one of the company's key brand names rather than a separate entity.

In February 1996, Rockwell Automation's Control and Information Group (C&IG) decided to create a staff position to evaluate the re-engineering opportunities within its new-product-development processes. Unlike many companies, which do not launch re-engineering programs until there is a financial crisis or serious drop in profitability, Rockwell Automation was highly profitable, and it had what was widely recognized as an effective program for developing new products. "We have had integrated-product-development teams for about 12 years and they've flourished," said Robert Murphy, director of new-product-development processes. "Our product-development processes have been benchmarked and studied by a lot of groups."

Rather, management chose to re-engineer primarily because it recognized that the company was being carried by its strong mix of products and technology, and not by development. "When we looked honestly at the processes themselves," Murphy said, "we saw many inefficiencies. We knew that they would inevitably affect the bottom line."

The re-engineering project officially started when Murphy's position was created. The first thing he did was evaluate all of C&IG's product-development processes. "I spent six weeks interviewing close to 120 people in all of the businesses that make up C&IG to determine which practices were most effective in their processes. I found that each group did some things very well, but no group did everything well." During these interviews, Murphy also asked employees what one thing they would improve. Their answers helped identify such problems as routinely exceeding resource capacity.

In a re-engineering effort, a common source of failure is inadequate communication. Citing studies showing that lapses in communication created problems far out of proportion to what would be expected, C&IG took painstaking efforts to keep employees informed through house publications, e-mail, and other forums.

By April, the re-engineering team developed an overall plan to guide the re-engineering process and identify key initiatives. To get approval for this plan, they started with senior management. "Many of the measures in this process required people to swallow a lot of tradition and a lot of history," Murphy said. "I knew that I would need to get top-down buy-in, or there would be problems. If we had not engaged the entire senior staff, middle management, and development-team leaders, I believe that the effort might have been initially successful but would not be sustainable over time." Because of the number and scope of people who participated in defining initiatives and moving them forward, no major stakeholder could refuse to buy in during implementation.

Management staff and product-development-team leaders then fleshed out the details, which were completed last June. New-product-process breakthrough teams were then created, comprising members from all business units and functional areas. These teams got into much more detail to figure out processes, tools, and changes required to achieve the goals. Those teams completed their work by October.

Since then, almost all new processes defined by these breakthrough teams have been implemented. Early findings are encouraging: "Everyone in the corporation understands that new-product development is the key to Rockwell's future," Murphy said.

The biggest problem C&IG faced before re-engineering was that the group wasn't effectively accounting for the practical capacity of its product-development resources—including expenses, capital, and manpower. Much of this problem stemmed from C&IG's corporate culture: Whenever there was a proposal to develop a new product or technology, management would typically approve it with little regard for available resources.

"We were so product-focused that it was difficult to pass up opportunities to add new products to the portfolio," Murphy said. "If someone had an idea that was excellent from technological and market perspectives, management more often than not said yes. Whether or not the project fit into the current product portfolio or—more significantly—whether or not there was the capacity to develop it properly wasn't easily recognizable." Midlevel managers who knew that the goals weren't realistic lacked the numbers to back up their assertions.

As a result, on paper up to 150 percent of resources often were committed at any time. The extra projects would be completed but delayed, and impacts would be felt on other projects. Such overcommitments caused conflicts, as the major programs tended to get the bulk of the attention; less-visible programs—far greater in number—often received inadequate resources, which forced project managers to do their best with what they had. The capacity overflows also caused morale problems among engineers who sometimes felt like they were overburdened, leaving inadequate time and tools to get their jobs done.

The need was clear for a better-communicated, more disciplined portfolio that was planned around a real resource-capacity model. To find that mix, C&IG is implementing a software-based distributed project-management system. For more than 10 years, Rockwell project managers had been using software packages such as Microsoft Project, Qwiknet, and Timeline. According to Murphy, these programs required that resources be unique to a particular project. "Once you start to bounce activities off of a pool of resources," he said, "the software fails. It does not allow splitting resources among projects." Since most resources—like quality teams, advanced manufacturing engineering, and component engineering—are multiplexed, C&IG had to be able to account for them properly. If such resources are overcommitted, they can often become the primary bottlenecks to releasing new projects.

To help improve management, C&IG has selected new software, AutoPLAN from Digital Tools Inc. in Cupertino, Calif., that allows multiple projects to draw from a common resource pool. This software package is integrated with an internally developed, Oracle-based distributed project-management system that does trending and predictions based on AutoPLAN-generated schedules.

C&IG is now moving closer to quantifying what 100 percent of capacity actually means. Ultimately, the goal will be to plan routinely to use 70 to 75 percent of this capacity. The rest will be reserved for unforeseen situations and emergencies—like continuation engineering efforts—that may absorb significant resources.

The new software and methodology are in the early stages of implementation, but C&IG reports that the results so far are very favorable. Divisions will start with a list of new products that they want in their portfolios. Managers will have facts rather than rely on intuition or anecdotal information to show how realistic a plan is, and they will be able to modify the plan to accomplish their goals. "At our company, this is a landmark event," Murphy said, "because we always have had trouble saying no to new opportunities."

Another benefit of the new software is improved communication. "We are now able to communicate more clearly," said Mark Knebusch, manager of worldwide business planning. "Before, the language was inconsistent. When I'd ask for an update from several project managers, I'd get a combination of Word documents, Powerpoint presentations, and maybe an Excel spreadsheet or two. I could get an idea of each project's status, but the presentation was very inconsistent and comparisons between projects were difficult. Now, if we want to drill down into some detail, we can do it in a consistent way and have confidence that dates and resources are realistic. It's now also much simpler to evaluate what-if scenarios."

The day-to-day impact on engineers should be significant. They should no longer butt up against each other for resource demands. The projects that do get launched are analyzed within the portfolio and can be achieved. "They won't have to face the frustration of running out of resources or being overtaxed," Murphy said. "Everyone is beginning to recognize that every project has the same chance to succeed, whether it's a $500,000 project or a $5 million project."

In addition to allocating resources, the re-engineering effort has changed the way the projects are managed. Monitoring projects in a stage-gate/milestone fashion is nothing new, but C&IG had unofficially discontinued the practice about five years ago. Once a project had been launched, it continued without formal periodic reassessments.

Now, C&IG has established a system to review each project throughout its major development phases and consider whether it deserves to be continued, delayed, or canceled. C&IG's review procedure, the project-authorization-request process (PAR), has four stages. PAR-1 approves funds for the market analysis of an idea. Just about anyone with an idea can submit a PAR-1. Based on the results of the marketing analysis, management can decide whether to fund the next phase.

If the idea passes through the first phase, its next milestone is PAR-2, the feasibility and engineering requirements phase. The phase's major activities are defining the engineering requirements—including the schedule plus quantified design and manufacturing resources—and developing the detailed project plan.

The implementation phase, PAR-3, is where the actual design and manufacturing take place. This is where most of the resources are expended.

Finally, PAR-4 is a project postmortem, which takes place about six months after the product becomes available for shipment. In this phase, people involved with the project answer questions regarding such items as how closely actual costs correlated with those in the budget. The information goes into an Oracle relational database that enables anyone working on future projects to benefit from the experiences. In the past, engineers and managers had to rely on anecdotes or institutional memory, which sometimes were unavailable or inaccurate.

C&IG manages resources through what it refers to as a hard-allocation/soft-allocation approach. At the start of a project, only resources for PAR-1 are hard-allocated, which means that the money actually has been transferred and is available to the project manager. As a result, a firm commitment is given to do that phase only. Money is hard-allocated for each subsequent phase only after approval.

However, right from the start, funds for the entire project are held in reserve in a theoretical pool, or soft-allocated. "If you're exploring multiple projects simultaneously and they all turn into viable initiatives," Murphy said, "resource-capacity problems could potentially develop. By soft-allocating funds, we ensure that if they do become viable the capital and manpower will be there. If the project is not approved, the soft-allocated money gets freed up and becomes available for other projects."

One relatively uncommon aspect of the re-engineered approval process is that when the given PAR phase is approved, all funds required to perform that phase are under the complete control of the project manager. Traditionally, when it came time for a capital investment, the project manager had to obtain the same 10 to 12 signatures that had been needed to get approval for the project in the first place. "Essentially, he or she had to fight to do something that was already given permission for," Murphy said. "This added no value. Re-evaluating the project at the end of each PAR phase does add value." At most, six signatures are required for each PAR phase, and all subordinate expenses for each phase are at the project manager's discretion.

Although Rockwell Automation's other divisions have not yet dedicated an independent, focused role similar to Murphy's, they're closely watching and monitoring what C&IG is doing. Once the group implements each measure and evaluates its strengths and weaknesses, it will pass on the finished product to the other divisions, so the benefits will be spread throughout the entire company.

home | features | weekly news | marketplace | departments | about ME | back issues | ASME | site search

© 1997 by The American Society of Mechanical Engineers