When UE, my former company, revised it’s idea system to recognize “small” ideas, there were many entries from production employees suggesting part simplification. Someone would note, for example, two screws differing in length by 1/64”. The longer screw could be used everywhere the shorter was required. Or two screws, one with a slotted head and one with a Philips head, were used on the same product. These “small” ideas are obvious to the folks on the shop floor but hidden to everyone else.
As the number of these idea submissions grew, engineers began to complain: “Drawing changes are expensive and time consuming, and the parts only cost a penny anyway.” Drawing changes are a topic for another post. But the pertinent misunderstanding regarding the part was that low cost parts, so-called C-items, like fasteners did not merit attention. Nevertheless, once the concept of part simplification was unleashed, ideas flowed in for nearly every commodity: labels, resistors, o-rings, switches and all manner of machined parts. On the floor, this part simplification also simplified product assembly (fewer tools, fewer stockouts, fewer storage locations), but detractors pointed out the relatively nominal impact on inventory levels for these penny parts. The problem with conventional cost accounting is that it focuses only on functional cost.
In 1989 during a Shingo Prize site visit to my plant (the very first visit for the Prize), George Koenigsaeker dropped a recently published book in my lap entitled “Variety Reduction Process.” I recall that he may have said something like, “This is an idea whose time has not yet come, but it may be useful to you.” He was right on both counts: First the time for variety reduction process (VRP) has apparently still not come, twenty years after the publication date of the book. In fact, I think the book is out of print, and there are few Internet references to the topic. Second, while the text describing it was cryptic and incomplete, VRP was extremely useful because it placed the many part simplification ideas we were implementing into a systematic framework.
An adequate description of VRP would violate my 5 minute read rule, so I’ll provide a Reader’s Digest version. According to VRP, there are two additional costs beyond functional cost that should be attributed to parts and processes. The first, Variety Cost, describes the cost arising from multiple mousetraps, i.e., multiple similar versions of the same part or process that provide the same function. The second, Control Cost, describes the overhead-related activity (things like drawings, inspection, ordering) to maintain excess variety. Excess variety may arise from value engineering of a part. For example, four different machined parts with identical dimensions: one is made from stainless steel, another from aluminum, one from brass and yet another from brass that is then nickel-plated. Each part may be designed to optimize the functional cost of the part, trading piece price against robustness. Consider for a moment the number of machine set-ups and other supporting activities to build and stock these parts. Yet cost accounting models ignore variety and control costs. In fact there are many other causes of needless variety that are never questioned, because the accumulated costs of this condition are not measured.
Within a given commodity, variety may expand unbridled without detection. VRP systematically exposes the condition. For example, using the VRP process at my former company, we identified twenty-seven different O-rings used in our products. Specific VRP techniques reduced the actual number needed to five. In another example, the number of different lead-wire assemblies was reduced by more than half, just by changing the product specification from “12” lead-wire outside enclosure” to “at least 12” lead-wire outside enclosure.” This variety reduction was transparent to the customer. Functional costs for individual parts sometimes went up slightly, but sometimes were also reduced through true economies of scale. However, the bottom-line for all product sales also magically rose. I like to say that for the parts we eliminated we had “reduced the EOQ, economical order quantity, to zero. “
How many parts or processes in your workplace can you think of that could have an EOQ=0? This thinking applies to processes and machines as well as parts. And the administrative opportunity can be as great as in production. Take a look around.
O.L.D
Your posts are great. They stimulate thought around the need to reduce complexity across product platforms. This post hits home where the proliferation of hardware is at uncontrollable levels. As more electronics companies outsource the design to multiple vendors where each vendor will introduce new hardware in every generation to maintain margins. The factories that produce the final configuration are buried under the burden of stocking parts for a diverse portfolio. How did you convince people beyond functional cost savings to undertake such initiatives? Are there rules of thumb on the cost to manage each part number?
Tony,
Thanks for your comments and great questions. Particularly in your industry, these are not questions with easy answers. I’ll offer my two cents. Perhaps other readers will also wish to chime in.
Convincing People. An EE friend told me once that any circuit could be designed using no more that three resistor values if that constraint were given to designers. This is not something I actually ever saw in practice however. Seemed like resistors were used more like bushings to make small adjustments here and there. ☺ My experience with “convincing people” of variety and control costs is that in a small company environment, cost accountants and managers are close enough to the floor to make balanced decisions. Also, engineers discovered after a short time, that simple improvements to the design process made their jobs easier as well: Easier retrieval of information about existing designs and better classification of parts are part of VRP. You’d get a few more ideas from the book.
Rules of thumb. As for your situation as a third party producer, outsourcing unfortunately has taken “out of sight, out of mind” to a whole new level. Add to this the rate of part obsolescence for electronic parts, and you’ve got a big challenge. You have limited opportunities to interact with your real customer not the one placing and expediting the purchase orders. If the products you are producing are new there may be a strategic opportunity for you and your customer to work together to balance apparent value engineering savings against standardization (i.e., few parts in your part master with commonality to multiple customers.) You should approach these customers and bring them to the floor to make your case. An increasing number of customers are becoming interested in just-in-time providers (i.e., able to build to their requirements with minimal stocks), not just-in-time suppliers (i.e., fat stockrooms to create a fast apparent lead-time.) For legacy products however, the opportunity for improvement is limited by the level of interest your customer has in engineering change – usually not much.
Regarding “rules of thumb”, once again the book will provide more detail. For example, technology trends are an important part of VRP. In the 1980’s (the old through-hole days), as designers spec’d in more 1% tolerance resistors for new designs, we discovered that our insertion equipment was accumulating double parts, complicating changeovers and adding considerable variety and control cost. A decision was made to standardize on 1% tolerance because 5% use in our products was declining. Using the VRP book for background, you could certainly develop rules of thumb for trade-offs.
There is more discussion we could have around this topic, but perhaps we can continue that offline at Oldleandude@aol.com in order not to bore folks who are not in your industry.
Bruce
I spent a few years working for a major electronic maker. For awhile, part of my job was to break down the Bill of Material (BOM) and look at the number of unique components required for the product. I then compared this to the number of unique components that our manufacturing sites could produce without changing over the chip placement machines. I had to go back to the design teams and ask them to change capacitors, resistors, etc… so that we could get the unique components under a certain count. This is how we tried to reduce the variation in components.
The bigger root cause of having my job has the design teams for the different sections of the circuit board sat in different areas of the building and very rarely every spoke with each other. That is different problem to solve though.
Matt,
Thanks for sharing your experience. I’ll borrow quote from a friend of mine, Preston Smith, who wrote a good book entitled, Developing Products in Half the Time: “The communication between engineers is inversely proportional to the square of the distance between them.” 🙂 In fact, another key idea from the VRP book is that “Production structure follows product structure.” So the problem of different locations for different disciplines may be yet another outcome of variety.
Bruce
Pingback: Il meglio della blogosfera lean #79 — Encob Blog
Pingback: Valu Ngineering | Old Lean Dude
Pingback: Silver Toaster Spirit | Old Lean Dude