Back in the early 1980's as we were participating in the formulation of an MRP II system for one of the of the major software houses, we were excited. ....after all, this was slated to be a package of packages, providing most of the features that we wanted at least in the current MRP circles. At the time, most MRP applications were custom-designed and custom-programmed, simply because standard packages weren't available. And...most of the support systems at the time were standalone, batch, and the idea of having a fully integrated real time application was promising.

Material Requirements Planning had been around since the mid-1970s and was reaching its heyday, receiving a lot of hype. Computerization made it possible to crunch through thousands of records in an effort to provide a more effective way in managing materials. Materials and MIS managers were embracing MRP, and championing its implementation to assist in achieving primary objectives: improving customer service, inventory investment, and plant operating efficiency. It would, according to the proponents, help us plan priorities for the shop, determine when and how much material to order.

We saw the vision of a system capable of helping us manage our inventories: move materials through the plant, keep track of parts, so that we would know....the status of a job,.... what parts were missing, ....when they would arrive,....what was on the shelf, etc. The emphasis was on control, and our new system would provide the means. Little did we know what we were in for.

A Trojan Horse

What followed was typical to the experiences of many in implementing MRP. The package didn't do what we thought it would; the next version of the software wasn't anything like the one we were working with. Phone call after phone call continued in attempting to get answers to technical questions....this didn't work,... that didn't work. The documentation was incomplete (or non-existant). We felt like we experienced the toils of war throughout the endeavor. The project schedule lengthened continually, and the budget was overspent, but we stayed with it determined to win the battle. Typically to most purchasers of this revolutionary product, we justified the package through inventory reductions.

But when the software was installed, inventories weren't reducing, and we began to discover the reasons. First we had to have ninety-five percent accuracy in our bills of material for the system to work. And..98% accuracy in our inventory balances. Then, we needed to "close the loop", to provide feedback to the planning process. And we needed a heavy dose of education and training to ingrain the new philosophy in the masses. Then we had to diligently deexpedite the materials we didn't need (and most of us did not under fire), or inventory would rise.....and it did. Customer service didn't improve, and we certainly didn't gain better control of operations. It seemed that the people who had the most to gain were the ones with a vested interest in MRP, particularly the educators and the software houses.

A Fundamental Weakness in our Plan

Today, MRP horror stories are so common, this one seems like ancient history. Article after article has been written on the merits of MRP, MRP vs. JIT, etc. Certainly it's easier to review our experiences in hindsight, and more difficult to anticipate while facing the unknown, but hopefully we can learn from hard experiences. It took a long time for us to realize how far off course we were. Instead of focusing in on the real causes of our problems, we were guilty of treating the symptoms: trying to control an unwieldy manufacturing organization.

It's tough to reason a war when you're in the foxhole, and bullets are flying over your head. But today we can look back and contemplate our mistakes. The lesson that we learned is fairly simple in this particular case: we were going all out to try and control an unwieldy factory, a near impossible task. Said differently, automation cannot compensate for decades of problems on the factory floor caused by haphazard growth. It is just too difficult, and requires too much effort to maintain it.

If we step back and review what's happened over the last thirty to forty years, we begin to realize that with few exceptions, our plants evolved by placing machines and equipment wherever space could be found. Our information systems developed into "islands", beginning with accounting, inventory, purchasing, etc., with no overall master plan to guide their interaction. We structured our people around functions, placing them into classes, separated them by walls, and then tried to manage this with MRP.

Consider that MRP assumes a finely tuned factory, which it almost never is, and requires too much effort to keep up with the dynamics of the shop floor, or the outsourcing of materials. And MRP programs are notorious for hogging computer time; they're generally run after hours. As a result, the gross/net requirement calculations are always out of sync with the real requirements in the factory. It releases orders according to lead times devised to represent a floor of disorder,...and it is not equipped to recover quickly from mistakes that occur, or deliveries that are late. The lot sizes, usually based on EOQ logic have little to do with real demand.

Most companies that we interface with seldom get through more than 60% of the MRP review reports weekly. Because the extensive amount of variances aren't maintained, over time, the effect is usually a build-up of unneeded materials, a consequential rise in inventory, and continued daily anguish over missing parts.

Deja Vu

Most companies haven't learned from the high washout rate that MRP has produced. Companies continue to pursue MRP implementations without regard to the real causes of their problems. A survey by Ingersoll Engineers revealed that companies with the least experience in extensive systems implementations are most likely to forge ahead with complex integrated_manufacturing support systems such as ERP.

Further, the same survey of manufacturing executives revealed that a surprising number of computer systems implemented have no links to operating profits. Most of the benefits that reportedly resulted from systems implementation centered around control. However, ironically, the largest share of those who reported successes with those results, against their original expectations listed their barometers as "casual observations." There's more.

Respondents reported a 30% overall (not just MRP) systems implementation failure rate, a documented testimonial to the difficulty of implementing manufacturing support systems. This wasn't a surprise, because of the 60-80% overall MRP failure rates reported in the past. So if implementations are so risky, with little real impact on profits, why continue to put blind faith and money into MRP? For control? GIVE ME A BREAK!!!

The Score Card is In

Today, software houses are in a slump, as is the rest of the computer industry. Companies are retrenching from continued investments in automation while they are trying to make their systems work. MRP implementations continue to run a high risk of failure, while direct profit producing benefits are "missing-in-action."

Because of the tainted history, we seriously question the validity of any claim to a net improvement to operating profit as a direct and sole result of implementing an MRP system. Far too many of the inventory reductions claimed to be a result of MRP could have been accomplished without the expensive system. However, few materials and MIS managers are willing to admit they've made multi-million dollar mistakes.

The irony is that companies are still managing themselves using old manual techniques, while automated systems are are pouring out data in massive volumes. We calculated that one company produced over 25 tons of unread MRP reports annually. Another produced three pounds of paper for every pound of production aircraft that left the runway.

Regrouping for an Offensive

If U.S. manufacturing is serious about regaining world market share in the 21st century, and we think it is, then our focus has to change, as well as the manner that we have been attacking problems in bits and pieces in search of a quick solution. We need to step back and refocus our efforts more broadly on what it will take to be seriously competitive in the next decade: the overall ability of the enterprise to respond quickly to the needs of the market, in new product introduction, in delivery, in quality, and in cost.

Organizations must be market driven, with focus on quickly satisfying the service chain; that chain of events that occur from the time that a customer inquires about an order, through complete satisfaction of the order: distribution, assembly, manufacture, and supply. All of the physical events must be enacted swiftly, accurately, and effectively. The faster parts, information, and decisions flow through an organization, the faster it can respond to customer orders. The keys are flow and time.

To make our factories highly responsive to the demands of the market, they must be streamlined: processes must be integrated, setups must be reduced using a zero-based goal, and all production interruptions eliminated. We must physically couple successive operations in the chain of work, removing nonvalue adding functions, and inducing velocity. As parts move quickly through the factory, the need to accumulate inventories declines, and the easiest way to manage inventories is to get rid of them.

Once this is accomplished, the factory floor can be managed with ease, and without the use of extensive, and complex material requirement planning routines. We have found that 60 to 70 percent of the benefits from streamlining flow and reducing cycle times can be achieved without heavy investment in technology.

While we have experienced a great amount of difficulty installing and effectively using MRP, it provided us with a good tool: the integrated manufacturing system, which uses the same sets of data, reducing redundancy and duplication. Using this as the center of the information base will take us a long way in organizing our information, making it timely, and more accurate.

But the flow of information must also be streamlined and electronically linked, so that the flow is direct--without interruptions and delays--eliminating queues. Business cycle times must be reduced to the time it actually takes to efficiently process information, supporting the quick movements of parts in the factory.

Having the ability to produce spontaneously upon demand requires an organization that is quick and resourceful. It requires short lines of communication, and velocity throughout the work chain. Every member must be in tune with the overall needs of the market, and close enough to one another to be spontaneous in helping each other support the common mission: serving the customer. Team play is a basic necessity to produce the type of results required to be competitive in the next decade.

This means we must restructure our organizations to induce effective and fast communications through out the enterprise: removing vertical and horizontal layers of bureaucracy in the organization chart, such that our companies become lean and agile, and decision time swift. Physical walls will have to come down. Our emphasis needs to be around the tasks that need to be performed to quickly satisfy a customer need.

The Next Campaign

If we are going to win the economic war, we are going to have to stop treating symptoms and begin to attack causes. We will have to quit trying to find a quick solution with bits and pieces of changes that have little overall effect on the profitability of our enterprises. We are going to have to make broad, sweeping changes in the way we operate, focusing on quickly serving the demands of the market, and managing our factories with ease.

Those companies who successfully transform themselves to accomplish this will be the winners and leaders: quick, resourceful, and agile. They are the ones who will have regained lost world market share, and with them, MRP, in turn, will seem like ancient history.

Please see our Agile Manufacturing Success Story (Click here)

Author

Richard G. Ligus is President of Rockford Consulting Group, Ltd., located in Rockford, IL., with over 30 years experience in manufacturing, procurement, transportation and distribution. He specializes in developing and implementing manufacturing, distribution, and supply chain strategies. Rich is an author and a speaker, and has developed seminars with the American Management Association. He is certified by both the Institute of Management Consultants and the The National Bureau of Certified Consultants.