Value stream mapping is gaining acceptance as a valuable tool for manufacturers of all types. This is because it is simple to use, yet it will accurately depict the relationship between value-added time (represented as the sum total of the cycle times of all processes in the value stream) and process waste (represented as the sum total of all inventory in the value stream). Because this column cannot do justice to all of the concepts and techniques associated with value stream mapping, I suggest exploring Web-based literature on the topic or reading the book “Learning to See” by Mike Rother and John Shook. The purpose of this month’s column is to share with you things companies are learning about their operations when they construct a value stream map. I like to call these “oh wow!” revelations. Such revelations are driving the changes needed to improve competitiveness in many organizations.
When developing its first value stream map, one company discovered that even though its main product required less than 1 hour of labor to produce, there was the equivalent of 1 year’s worth of inventory in the pipeline (with a majority of this being work-in-process). Simply put, if the company started to produce this product today, it would not be finished for almost a year. Because this amount of inventory was a symptom of problems in the process, the company set out to find the root causes and develop solutions.
Another company learned that paperwork was one of the major causes of delay. As a rule, the paperwork was either not completed on time or not completed accurately. As a result, production personnel were wasting time trying to get paperwork issues resolved instead of running machines.
In creating a value stream map, another company saw the impact of batch production quantities that did not match customer ordering quantities. Customers were buying in either small or large tote quantities. The selected production batch quantities were too small for the large customer order quantities, yet they were too large for the small customer order quantities. The result was that the company had insufficient inventory to fulfill large tote orders, or it had too many parts left over after fulfilling small tote orders.
During another company’s value stream mapping event, management realized that the overall process yield was much lower than anyone had anticipated. In this company, there were five sequential production processes. Taken individually, process yields did not look that bad—they ranged from a low of 95 percent to a high of 99 percent. But the overall process yield, which is derived by multiplying (not adding) individual process yields, was actually 86 percent. Another company, upon completion of its value stream map, expressed its quality problems another way. As a result of frequent rework requirements, it actually took more than 3,800 hours to complete 2,700 hours of labor. Put another way, a quarter of the company’s effort was devoted to doing things over.
A value stream map revealed something interesting about how one company was purchasing raw materials. Although this company had negotiated twice weekly deliveries of a certain raw material, it was carrying more than 3 weeks of inventory of that material (expressed in days of customer demand, not production capacity), and the material was still coming in on a regularly scheduled basis.
Another company realized that its high level of machining efficiency was killing it. The company had worked hard to program its CNC machines to produce different parts in one setup. For every machine cycle, one each of three different parts was completed. When this was initially thought out, demand for the three parts was equal, but over time, demand on one of the parts had decreased significantly. The value stream map highlighted 9 months of inventory for one part, while the other two parts maintained a more reasonable 4-week inventory level. Finally, a value stream mapping event confirmed what one company instinctively knew, but had not acted upon. There was a huge work-in-process inventory buildup immediately after the machine with the longest setup time. As would be expected, large batch sizes are generally run on machines with long setup times, based on the idea that once the machine is set up, it should be kept running. Unfortunately, the operations that followed this machine never seemed to have the right parts (although they did have an abundance of the wrong parts). It was readily apparent that batch sizes on this machine had to be reduced so that the right parts would be available when needed.
Try completing a value stream map for one of your products. You, too, may experience an “oh wow!” revelation.