Making improvements using lean principles in a low volume/high mix (LVHM) manufacturing environment can be difficult, especially when trying to determine where to start. However, some of the same tools that can be applied in traditional mass-production environments work equally well in a high-mix environment.
As the engineering manager for a LVHM organization in the medical device field, one of the keys to success for us was to identify groups of products by flow – especially during one project I worked on to fight long lead times and noncompetitive costs. In a sense, we stopped trying to compare apples to oranges and just looked at our products equally. We would first work to identify all the products in a facility, department, or office and then define the flows. This was probably the most time-consuming task for us as some facilities had as many as 4,000 products – some of which were only made once per year or less in as low a quantity as five (excluding custom orders and prototypes).
We would then align the products by flow. For example, in one manufacturing facility, instead of having products grouped by hips, knees and shoulders, we would instead group them by “Lathe-Vertical Mill-Vertical Mill” or “Lathe-Mill/Turn-Mill” or “laser etch-non sterile packaging – labeling” process flows. They might have been apples vs. oranges vs. pears in appearance, but if they all used the same flow, that was our starting point.
After grouping the product-by-product flows, we could then start identifying the volume, takt, and capacity required for each flow. Sometimes this meant products A, C and X would be grouped together, even though traditionally they were considered unrelated and ultimately look nothing alike. This created smaller individualized value streams within the overall organization – some with 300 products, others with 50, and still others with just one or two. Once we understood this, we then knew how many manufacturing “units” (e.g. machines, secondary processing, labor, etc.) we would need for each value stream. This fed our capacity planning. Ultimately our VSM ended up looking something the image in the header above.
This overall strategic-planning work allows you to determine where to start value-stream mapping, and what percentage of your business each flow represents. If you add current margins and variance reporting (if you have the data) of the combined value stream and detail by each smaller product family within the stream, you can start making some intelligent choices on which VSM will probably yield more significant wins for the organization.
For example, we focused on one combined value stream that represented several hundred products and reduced the average changeover time from 10 hours to just 30 minutes. Not quite single minute, but still a 95 percent reduction in wasted time between orders that meant the difference between losing money and making a decent profit on the entire range of products.
We started seeing significant growth once we had defined and improved 50 percent of our revenue using this strategy. We targeted one year for this initial part of the transformation after figuring out how many “events” we would need to get these changes done AND get a large percentage of our staff involved in the cultural change as well. And things only got better from there – once we hit 80 percent of our revenue, the sky was the limit.
I challenge you to put aside some time to intentionally consider your organization’s value streams – or your product families if you don’t think in terms of value streams – and draw out your flows. You might find hidden opportunities within your current assumptions, or perhaps you’ll just find a starting point for transforming your LVHM business.
Have you already had success defining and improving value streams in an LVHM environment? Comment about your experience below to help others on their lean journey.
I assume that you would need the cycle times first before creating the VSM for it?
What if the cycle times are not available as in a LVHM environment this would be very challenging?
Isn‘t this the classical Part-Quantity-Process-Routing (PQPR) approach?