Process Is Back

Software alone may not yield the best results for manufacturing.

Half a dozen years ago, business reengineering was hot as a slab of continuous-cast steel. It seemed every business executive was reading Reengineering the Corporation (1993, HarperBusiness), Michael Hammer and James Champy's best-selling book on the topic, if not mandating some major process redesign effort in the company. Then reengineering -- often called business-process reengineering or BPR -- suddenly got a black eye. Speaking at a conference, Hammer himself bristled when a questioner asked him if, in fact, reengineering wasn't merely a fancy word for layoffs. Hammer, of course, disagreed. But the knockout punch had been landed, and many companies that had once embraced reengineering suddenly came down with a case of BPR amnesia. They jumped instead to the "next big thing" which was -- you guessed it -- enterprise-resource planning (ERP). Although ERP software promised a great deal, many manufacturing companies derived little, if any, significant gains in their processes from the new systems. The reason is that SAP AG's systems typically were designed for corporate operations such as financial management and order taking, processes that offer limited functionality for manufacturing. "Even though SAP was born in a manufacturing environment, they to date are unable to really develop a robust set of functionality to handle all the issues manufacturing organizations need," says Russell Brackett, senior partner in Computer Sciences Corp.'s CSC Consulting Inc. Indeed, the obsession American corporations have had in recent years with expensive and complex software programs -- often purchased to avoid changing all their older systems to sidestep the ubiquitous Year 2000 computer-date-change problem -- may have gotten out of hand, some observers say. "People are focusing on systems work and not on the processes underlying them," Brackett observes. "The outcome of that was to let technology lead." Adds Buzz Adams, president of Peak Value Consulting Inc., "There's a belief that the new systems will automatically reengineer your processes. Most people put in new systems with the hope that they will improve productivity and cut costs." To their dismay, some companies that installed ERP systems found little payoff in terms of process gains. Part of the reason, Brackett says, is that companies were using technology "to pave the cow path. They used it to automate the way things are done today," he says. "Companies like to throw money at technology solutions, and nine out of 10 don't work out," observes Stephen McMahon, formerly director of lean-manufacturing practice at Coleman Consulting Group and now solutions manager with IBM Corp.'s Global Services unit. "Technology has been relied upon to do a lot of what processes should do. Sometimes the simplest nontechnology solution is the easiest and most efficient way to go." A potential downside of too much dependence on systems is that manufacturers can become overly dependent on technology. Ironically, waste and other inefficiencies in processes -- in the office or on the plant floor -- may be the root of the pain they sought to cure with the software in the first place. Now it looks as if the pendulum is starting to swing the other way. "We've had more calls in the last six to nine months for business-process reengineering than we had in the previous three years," says Brackett. "These companies have implemented the technology, and now they want to go back and get the process benefit. Many of them are in for a pretty shocking realization, because it isn't easy to do that. Once you've poured cement around those old processes, it's very difficult to change them." Brackett cites as an example an aerospace-industry supplier that had been laboring with an older substandard material-requirements-planning (MRP) system. The manufacturer was able to eliminate 50% of the work in process through a variety of kaizen-type process changes before attempting to upgrade its systems. A classic example of a manufacturer that wisely improved its processes first is Caterpillar Inc., Peoria, Ill. The heavy-equipment manufacturer's now-legendary Plant With A Future (PWAF) project in the early 1990s focused the organization on radical improvements in its manufacturing processes to boost competitiveness. The architect of PWAF, former Caterpillar group president Pierre Guerindon, stated that "automating complexity is merely mechanizing waste." He believed in the notion of taking a "hard look at products and processes -- and simplifying both." The end result for Caterpillar was, in the words of process champion Guerindon, "products and processes that were simple enough to automate cost-effectively." Certainly, manufacturers that have embraced the concepts of lean manufacturing have found significant gains can be achieved from putting process issues first and technology second. And the process-first approach applies to corporate functions such as accounting and human resources as well. Few experts would disagree that processes are more important than the computer systems that support them. As such, they agree, processes should be fixed before installing any software to automate them. "But once you get the processes straightened out, new systems can push you to a new level," Adams admits. McMahon agrees, adding, "If you're simply going to implement a just-in-time (JIT) or pull system without any technology to add discipline to the process, it's going to fail just as badly." Sometimes organizational issues are the root cause of inefficiencies. For instance, before Elf Atochem North America overhauled its business processes and installed a new enterprise-wide SAP R/3 system, processes tended to be fragmented among each of a dozen business units. Each, for example, had handled customers differently. That meant customers had to deal with multiple units of the company just to place a single order. Likewise, payment for the same order required several invoices. Processing of the order inside the company was equally diffused and time consuming, with several units getting involved. The separation of inventory and production among different units meant that no single unit could manage these functions. Millions of dollars of inventory were written off annually. Of course, with separate ordering and production systems, order takers were unable to promise firm delivery times to customers. The solution was consolidation of the company's accounts-receivable and credit departments so that customer orders could be aggregated into a single account with a single invoice. Customer-service activities also were combined so that customers would have one unit to deal with for order status and problem resolution. Sometimes, understanding how a process works requires executives to go into the plant or out in the field to see exactly what employees do and where the potential inefficiencies lie. For instance, when Adams was vice president of finance at McKessonHBOC Water Products Co., a Pasadena, Calif.-based division of McKessonHBOC Inc., he rode in the water-delivery trucks to learn the company's core business -- delivering water to half a million homes and businesses. A process guru with a finance background, Adams has built a consulting practice out of helping companies wring out wasted effort. At McKessonHBOC Water, he found that 87% of the tasks supporting the company's route-accounting process were nonvalue activities. Eliminating them saved the company $1 million, and reduced the route-accounting staff by 90 positions. Inadequate route-accounting controls also added up to missing inventory. Drivers would record on paper the number of bottles delivered and what remained in the truck and turn the report into the manager each day. Discrepancies, though, were rampant, and the company had no way of checking what was accurate. What's more, Adams says, "The check-in and -out was awkward," wasting still more time. "They didn't need to be doing double-checking of inventory." In that instance, technology played a part in fixing the process, as workers were equipped with handheld devices that use radio-frequency telecommunications to connect with the company's database. Also, poor design of the trucks was causing the drivers to spend far more time jockeying around with full bottles and returned empties than was really necessary. Not only were the trucks redesigned to speed on- and off-loading, but more careful management of bottles and coolers used by customers helped the company save more than $8 million annually, Adams reports. He recommends that companies looking to improve processes start by assessing their overall strategy and then carefully analyze their processes to see that they support it efficiently. "Most large companies don't understand why they are unique," he says. "As a result, there aren't very many large companies that are process-smart." One reason is that large companies tend to be made up of several departments, each with its own systems. "Since a lot of systems have been built based on department needs, they're not geared to a process," Adams argues. A process often spans several departmental units such as accounting, sales, production, and shipping. Another reason most systems are inherently wasteful is unnecessary complexity. "Most systems are unnecessarily complex because companies repaved the manual road that was there with technology. But if you have simple processes, your systems will work a lot better, too," Adams advises. That complexity requires a careful breakdown of each process in order to understand it. "For most people, it's very difficult to understand all the steps and complexity of a process," he adds. Analysis of a process such as order taking, credit checking, or packing and shipping requires a detailed walk-through of each task comprising the entire activity. Next, Adams advises, questionnaires should be sent to the supervisor of each department, with the aim of obtaining a detailed accounting for people and time, down to the task level. Some consulting firms will do this, Adams says, but they'll charge more because they often do the information gathering through in-person interviews with supervisors and employees. Other information-department managers should be asked what major processes they perform daily, weekly, monthly, quarterly, and annually. Then each process should be broken down into tasks. Who performs those tasks should be delineated as well. Finally, Adams recommends, all this information should be either put into a software program for mapping or laid out on a wall using, say, sticky notes. An automated alternative is to use a software package from a company such as Phios Corp., which helps manage knowledge contained in some 5,000 common business processes. Dow Corning Corp. is using the Phios system to document all its process designs and make the information available throughout the company. The next step is what Adams calls a value analysis: Each task gets graded on whether it represents value-added or non-value-added activity. Supervisors should be involved in this process, he says, which by itself could take up to five days. "What you find, typically, is that 65% to 70% of the tasks are nonvalue [tasks]," Adams says. "At the lowest level of detail, you find massive amounts of waste." For instance, he says, tasks such as fixing errors caused in earlier steps in the process are a dead giveaway. "If you find you're doing rework, it's nonvalue effort." Identifying wasted effort is the hardest part for most companies. "Deciding what's causing the nonvalue effort is the hardest part," Adams says. "Once you figure this out, the solutions actually are very easy." In many instances, the root cause of wasted effort lies in inadequate employee training. Employees often aren't instructed in the correct manner to perform a task, so they end up duplicating someone else's mistakes or creating their own. But training people to look for and eliminate wasted effort can pay off. "Once you've trained people to understand value versus nonvalue," Adams says, "they will apply it forever in all their work. And it takes just a couple days to do it." He says employees should be encouraged to ask questions such as, "Why am I checking things?" or "Who is making the errors?" The source of a process problem occasionally can be traced to something as simple as an inefficient or costly rule that governs it. At McKessonHBOC General Medical, in Richmond, Va., the company had been losing a 2% discount it could have earned on bills if it processed and paid them within a 20-day period. Part of the reason was that many invoices had to be kicked back for additional manual handling because of a $50 tolerance level for discrepancies. The policy of the distributor of medical and surgical supplies to hospitals and clinics was not to issue immediate payment on an invoice if there was an error of $50 or more. Thus, a $100,000 invoice couldn't be paid because of a $75 error, holding up its processing for days and missing the discount. With millions of dollars worth of invoices being processed weekly, the company was losing $200,000 each week in potential discounts simply because of a rule that was saving a minimal amount of money. "It took eight days of programming time to change it," Adams reports. Fixing the process yielded $10 million a year in discounts. Another reason was a bottleneck in the mailroom, where 11 people sorted huge volumes of mail in a way that slowed down the entire process for days. "Just by changing procedures in the mailroom, seven positions were eliminated and the process was sped up by one to two days," Adams says. In both instances, the gains were achieved by cutting time out of the process.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.