Many years ago I was in graduate school at Carnegie Mellon University, and signed up for the Ph.D. qualifying exams, “quals”, to be admitted to the program officially (I wrote about why I left the Ph.D. program a year later here). One of the topics available to be tested on was Manufacturing. This exam, for the most part, consisted of the panel holding up things and then asking you to come up with a way to make it. For the better part of two months after I signed up for the quals, everything I looked at was examined, considered, dissected, and mulled over… to the point where everyone I knew was sick and tired of my ruminations over salt shakers, paper clips, tires, and everything else I saw.
Binder Clips and More
So I stepped into the spotlight, and one of the panel members held up… a binder clip. I spent the next 30 minutes hypothesizing a way of making binder clips. As it turned out it was not the way – they enlightened me as to how they’re really made at the end of the session – but I’d come up with a viable method nonetheless. Since I still had 20 minutes, I was then quizzed on other things, including Lean Manufacturing.
One of the things “in the air” at the time was a discussion of Work-In-Progress, or WIP, and how there should – in theory – be zero WIP at any point in the line. The idea was that WIP would insulate bad processes from being found. Theory also stated that idle inventory that couldn’t be sold was not, in fact, an asset but a liability representing cash tied up (which I agree with). So, drive it to zero, let the problems come bubbling up, and then work to eliminate them. Fine and dandy, but doing so very likely would result in disruptions in the flow of products off the end of the line.
This prompted one professor to then ask “What about the customer’s POF?”
After asking what that acronym stood for, he said “Pissed Off Factor.” On the fly I discussed how the theory called for zero WIP, but the reality of most production these days was one where the customer worked to have replaceable suppliers – and so the supplier needed protection to ensure retaining the customer. My belief was that there should be some WIP and some end-of-line inventory to insulate the customer from any hiccups that might happen as the line was brought into a Lean state.
I passed the exam; but the cognitive clash between theory and practice led to my coining this quotation:
In theory, there is no difference between theory and practice. In practice, there is.
Fast Forward to an Application
While at Ford Motor Company in Sandusky, Ohio, I was tapped to be the initial Lean Champion in first bringing in Lean concepts to the plant. After doing some independent studying including reading the book The Machine That Changed The World, taking a course or three which included tours of several “showcase” companies willing to let people see what they were doing, and hearing the seminars by the Lean consultant hired by the company – a Professor from MIT who had won the Shingo Prize, among others – and who had developed a comprehensive set of Lean Guidelines for our use.
We were in the initial stages of designing the production lines for the Ford Focus headlight and, following hot on that, the newly-restyled Ford Taurus / Mercury Sable headlight. Although an initial line layout had already been done, I got permission to consider a total revamp. Thus, I got representatives from Industrial Engineering, the Equipment group, the New Model Launch Engineers, and so on, and locked them in a room with that MIT consultant for a one-day concentrated Lean course. The next day, with that consultant acting as a co-facilitator with me, we worked on trying to develop a new layout incorporating lessons from the seminar.
The result was astonishing. While I can’t, for confidentiality reasons, go into enormous detail, one key “earthquake” in line design philosophy made an enormous difference with multiple consequences. Until that moment, all lighting lines were set up with a two-shift philosophy: one shift would run right-hand lights, another shift would run left-hand lights. By definition, two shifts were required to send matched pairs to the assembly plants.
This earthquake-level shift was simple: two back-to-back lines running left and right hand lights simultaneously. Every pass through of the concept yielded further ripple effect improvements, in many respects coming from the functional doubling of TAKT time for each line. In the end, we saved over $1 million in capital equipment costs and reduced labor content by over $100K per year. Plus the duplication of the lines into back-to-back cells ensured capacity, even if reduced, if there were an issue with any of the critical equipment – an option not present in the “one big line” layout.
Square Pegs, Round Holes
One of the fundamental things taught was that in an ideal Lean line people act as the conveyor. A part would be taken out by a person, the next part put in, and the person would then carry the finished part to the next station. In the meantime, the machine would perform the operation and verify its own work. I came to admire the idea of autonomation, the term I learned for where the machine not only did the value-added actions required but then verified they had been done correctly. The idea being simple: zero defects passed on to the next station.
Because of the positioning requirements for several assembly operations, however, there remained an asynchronous pallet system. On top of that, some of the components, e.g., a wire spring, were exceedingly difficult to conceive of an automatic system to be able to install them. We did a cost-benefit analysis and it ended up being far, far more cost effective to have a stationary person working to do the component assembly than to develop a super-duper-and-by-definition-custom piece of equipment to accomplish the task. For these stations we did build in Poke Yoke capability instead of having the station downstream from the value-added operations; again, done under the idea of checking the work then-and-there, thus eliminating the risk of passing on a defect.
I made a pilgrimage to Dearborn to report to the company’s Lean Group about our progress. In attendance was that MIT consultant, whose reaction was totally opposite of what I expected. Showing enormous capital and labor savings, I expected our team to receive accolades; instead, that consultant blew up because we had not followed all of his guidelines to the letter.
Some people “got it” but some didn’t: Lean – like any discipline such as Theory of Constraints (which I still believe in whole-heartedly) – requires intelligent application, not mindless application of The Rules even when they don’t make sense.
As the line design and construction proceeded, I stepped aside as expertise spread and a more formal team formed. Comprised of representatives from across the company on a co-location assignment to watch and shepherd the development and implementation of the new lines and then take the knowledge back to their plants, they did an admirable job. And I was glad to be freed to pursue other projects. But I did still have one last contribution to make.
Theory said that parts should be packed and then loaded straight into the waiting truck. No WIP. But I asked “What can go wrong on the line?”, as in a serious disruption of production flow. The robot, which poured the glue joining the lens to the housing, was the most common piece of equipment that went down. “How long until it’s repaired?” Mean-Time-To-Repair was, IIRC, three hours. So I proposed we have a four-hour rolling stock of finished goods to protect us from such events.
From the shocked look of most of the team, and the horrified reaction of the MIT consultant who had come down to review the area, I must have been proposing we turn the entire plant into a flower farm. You could almost hear the scorn as they repeated the word: INVENTORY?
But there was a method to my apparent madness. Ford charged suppliers who didn’t deliver on time a staggering amount per minute that the parts were not there. Suppliers had been bankrupted that way. I argued that the carrying cost in both sunk cash, and space, was minimal compared to the consequences if we failed to deliver on time. I presented the math: the carrying cost of four hours of inventory vs. the financial consequences if we didn’t deliver on time. After all, it was not a matter of playing the odds. The robot would go down sooner or later, and without that inventory we would fail to deliver on time.
And the Winner Was…
After much debate, practicality won over theory. We had rolling inventory of finished goods at the end of the line.
- Lean is a great thing, but do it intelligently considering the uniqueness of your situation
- Rigid adherence to every guideline is a guarantee for failure
- Include a consideration of the POF in every decision
© 2014, David Hunt, PE