Bump and Grind

Forcing lean on a fundamentally broken system

Published: Wednesday, March 9, 2016 - 16:25

Here’s a personal reflection from my distant past that might describe a current state for some of you.

When I began working in manufacturing during the pre-lean era, the quoted lead time for my company’s products averaged 12 to 16 weeks. By the 1980s, however, many customers began to routinely object to our promised deliveries, and we would then try to accommodate them by moving “hot” jobs to the front of the queue and bumping the jobs that were already there. Bumping, of course, is like cutting in line except it’s not immediately obvious to those that are bumped.

Eventually the bumpees also become irritated. One long-time customer remarked to me, “I always give you plenty of lead time, but you still miss the due date.” This unfortunately was true—and inevitable. We bumped customers until they were late. To paraphrase W. Edwards Deming, our bump-and-grind production system was “perfectly designed” to produce long lead times, late deliveries, and dissatisfied customers. We tried mightily to mask our problem with tactical workarounds, but they were doomed to fail. Here’s a short list of perverse tactics we tried.

1. Visual management. As the number of customers expecting shorter lead times grew, it became apparent that we couldn’t bump the queue for everyone. So we prioritized the queue. Orders for export were stamped with a big red “F” (for foreign). Orders from new customers were marked with an “N,” and big customers got a “$.” Then there were orders that were deemed hot only because of external pressure. For example, a few customers sent expediters to sit in our lobby until their order was complete. These orders were marked with a colorful bull’s-eye. Regrettably, orders with no special mark were deemed unimportant, and due dates printed on them were largely ignored.

2. High-runner products were built to stock with a rationale of creating an apparent just-in-time supply to our customers. Unfortunately, factory orders for stock became fillers, to be completed after all the aforementioned hot jobs were finished. “We build for orders first, and then for stock,” a scheduler told me. Consequently, our pick rate from stock was less than 60 percent.

3. Daily huddle. Because of our new taxonomy of hot jobs (i.e., F, N, $), it became necessary to meet each morning to further stratify the hottest of the hot from the more tepid. This was a “least-worst choice” meeting. The reality is that when you can’t deliver, there are no good choices. We also covered part shortages during this huddle. We were always robbing from Peter to pay Paul, and as we jerked our schedule around in response to customer complaints, we inadvertently created shortages and expedites with suppliers. Managing stock-outs became the major activity for buyers. (This was also supported by a multicolored visual stock-out system.)

4. Concierge service. As a countermeasure to poor delivery performance, some customers placed orders well in advance of need to ensure they were in the queue. However, our FISH (first in, still here) scheduling essentially neutralized that ploy. More savvy customers realized the best path for service required a demand for very short lead times. This gave them visibility in the queue, but soon led to the dubious imposition of premium charges for fast delivery. In other words, “We can’t deliver in the time frame you want, but if you pay us more, then we can.” Some customers balked, but others just passed the increase on. From an accounting standpoint, this pay-to-play model seemed like a moneymaker, while in reality it was a reward for poor performance. (As a footnote, a few customers realized that if they needed an order for, say, 10 products, they could place two separate orders, one for one piece at the premium price and a second for nine pieces at the standard price, and all would be shipped on the premium date. They saw through our tactics.)

Altogether, our bump-and-grind production system created a high-stress environment for provider and consumer alike, a cat-and-mouse game where each system tweak by the provider missed the mark and created unanticipated reactions from the customer. Production scheduling became a juggling act, and we hired people who could tolerate that pressure. The painful lesson for me was that no amount of tactics can overcome a bad strategy. Fortunately, my company discovered the Toyota Production System, and everything changed. That was 30 years ago.

It’s incredible how hard we’ll try to tweak a system that is fundamentally wrong before we recognize the need for an overhaul.

How about you? Is your production system in need of an overhaul? Please share a story.

About The Author

Bruce Hamilton

Bruce Hamilton, president of the Greater Boston Manufacturing Partnership (GBMP), brings hands-on experience as a manager, teacher, and change agent. Prior to GBMP, Hamilton led efforts to transform United Electric Controls Co.’s production from a traditional batch factory to a single-piece-flow environment that has become an international showcase. Hamilton has spoken internationally on lean manufacturing, employee involvement, continuous improvement, and implementing change. Also, he has contributed to numerous texts ranging from visual control to variety reduction. Hamilton’s blog, Old Lean Dude, is an ongoing reflection on lean philosophy and practices, with an emphasis on keeping good jobs close to home.