Installment #4 continued from:
What impressed me most about my company’s strategy in introducing and managing change was the deductive (general-to-specific) approach we used to educate our clients and team members. It impressed me because previous Japanese bosses had always been inductive (specific-to-general) in their training and presentation methods, not at all forthcoming with big picture information.
So to my surprise, our approach wasn’t about us walking in as “experts” and telling clients what to do; instead we provided them a framework that gave context and meaning to all improvement activities that would follow.
My boss would begin his rap at the 20-thousand-foot big-picture level, so he could later swoop us all in for concrete analysis, process redesign and actual implementation.
Our starting point was defining “productivity.” We kept it simple:
Productivity is the ratio of input to output.
We explained the different ways to express the input-output ratio. One is input over output, for example, man-hours required (input) to make a single widget (output). Another way to express the ratio is to flip the equation over and look at output-over-input. An example would be number of widgets produced (output) per hour (input).
High input is generally considered bad, while high output is good. Conversely, low input is good, but low output is bad. This means efficiency-improvement kaizen activities strive to maximize output while minimizing input.
Hence we explained that the basic ways to achieve productivity improvement were to: 1) maintain same input with increased output; 2) reduce input with the same output; or ideally 3) reduce input and increase output.
Once our team members got their heads around the concept, we provided an overview describing the dimensions of productivity. We presented it as such:
Productivity = Method x Performance x Utilization
We would break down the three dimensions in a way that anyone could understand. We’d explain that method is “how” work is done; performance is the actually time it takes to complete work in relation to how long it’s supposed to take; and utilization is uptime, or “non idle” time of both equipment and workers.
We then expounded on each dimension.
As mentioned, method is related to how work is done. A concrete example tells the story:
Suppose an inefficient production line is set up to produce 50 parts per hour. Let’s also assume the production operator is taking lots of unnecessary steps to do her job, bending at the waste to pick up stacked material, reworking bad products coming off the production line, etc., all work that is unnecessary to making a good part, and therefore, wasteful. This fictional method is terrible because the input is high relative to the output.
Now let’s think about how we might improve the efficiency of such a line. Ideally we would analyze the current situation, brainstorm improvement ideas, and likely decide that a more compactly laid-out line would reduce unnecessary steps for the operator, that a spring-loaded table would keep line-side material at waist level to eliminate operator bending, that finding and eliminating root-causes for quality issues would provide stable processes that produced only quality parts and thus eliminate the need to rework, etc.
Extending the example, let’s make another assumption: that our new-and-improved redesigned process can (theoretically) produce 100 parts per hour. In this case our input remains the same (1 hour) but the output (number of parts produced) doubles. The result is the labor cost is halved, a substantial chunk of change, a magnitude of improvement that was not at all unusual in the projects we led.
Once our team members understood the “method” dimension of productivity, we let them know upfront that we planned to do lots of method improvements moving forward, but that we couldn’t do it without their help. And every word of that was true.
Performance and Invisible Culture Gaps
Next we would explain the “performance” dimension as the ratio of actual work time in relation to the time it is supposed to take, also known as “standard time.” So using our previous example, suppose standard time under this wasteful process translates to 50 parts per hour, but we are actually only making 40, then our performance to standard is only 80%. (40/50=80%)
This is where we normally let our clients know that performance improvement was a longer-term project than method improvements. And it’s also where I stumbled onto a huge gap in how Japanese and American managers frame the performance challenge.
In the land of individualism (and all those precise individual job descriptions that come with it), responsibility for performance improvement is clearly placed on each individual’s shoulders, from the president on down to the production worker. In simple factory terms, this means American operators are responsible for their own performance, in short, for simply working faster.
This is not at all how my Japanese superiors saw the world. Through their Confucian, collectivist lens, performance was driven by employee morale, and high morale was deemed only possible if leaders provided the necessary and sufficient conditions that nurtured high morale. (See The Physics of Responsibility.) That was the theory anyway. And I bought it unconditionally. (Unfortunately for me and my teammates, our boss’s morale theory didn’t apply to the work situation in our company.)
Precisely because performance was a long, drawn-out, leadership-development project—or perhaps because we weren’t good at the morale-building schtick anyway—we chose to put performance-related projects on the back-burner and instead focus on getting the quick impacts that come from method improvements, along with improving the last key dimension in the productivity formula, “Utilization.”
Utilization measures the “uptime” of both equipment and operators. In the ideal world you want the operators working in harmony with the equipment at the minimal machine cycle required to produce conforming product. So not only do you want operators in sync with one another, you want man and machine in sync as well, which basically means not bottlenecking each other. (Note that utilization goes out the window in the absence of customer orders, as the lean company without orders will always choose to shut down the line.)
Assuming customer orders are rolling in, let’s use some friendly round numbers to tell the story: if a machine is down 20% of the time for whatever reasons, the uptime—or utilization—is 80%. Aside from equipment problems, line stoppages due to quality issues, etc., in some cases the machine is down because it’s waiting for the operator(s) to finish the work.
And sometimes it’s the opposite—the operator finishes the work early and is waiting for the machine to finish cycling.
Either way they are not in sync, so one or the other is idle.
In the first example, you’d attack the machine downtime problem by figuring out reasons for stoppages, quantify downtime by category, then line up the categories on a Pareto chart. Next you’d work on figuring out the root causes for the top downtime offenders and implement corrective action accordingly to eliminate the causes and, consequently recurrence. All in a perfect world.
If, on the other hand, the operator is the bottleneck, then a method improvement is in order with a focus (perhaps) on reducing motion from unnecessary steps, bending from the waist, etc. Ideally the new method would be designed so an average-skilled operator could comfortably complete all tasks within the target machine cycle. Short of that, adding an operator to the process might also be a viable option depending on how the numbers crunched.
Once the Method x Performance x Utilization formula was understood, we’d plug in the numbers to tie it all back together again.
Getting back to our make-believe wasteful line that was producing 50 parts per hour, we also established that our new-and-improved design would theoretically get us 100 parts an hour. So the efficiency of our current method in relation to the new-and-improved one is 0.5, since it’s only making 50% of the parts it could make under our superior method.
As for the performance piece, if standard time under the current method calculates to 50 parts per hour, but the actual output is only 40, then the performance to standard is 0.8, because the process is only producing 80% of what it’s designed to produce.
Finally, utilization—if the line is down 20% of the time for whatever reason, it means the uptime is .8 or 80%.
Productivity = Method (0.5) x Performance (0.8) x. Utilization (0.8 )= 32%
So to summarize and simplify: for work to be productive, its method is efficiently designed (devoid of waste and unnecessary motion), the work done by highly motivated production workers dedicated to hitting production targets, running at maximum uptime in harmony with equipment and technology.
Interestingly when we’d apply our productivity equation to our clients’ factories, the American staff would freak out because the final number was so much lower than they expected. Most American clients equated “productivity” with only the “performance-to-standard” piece of the puzzle. So if you described the make-believe wasteful production line above, most of them would assume “productivity” is 80% since only performance-to-standard was on their radar. No surprise that when you hit them with a low number (like 32%) their feelings got hurt. In their own eyes, it made them look bad.
But surprise surprise, my Japanese comrades saw the world differently. Their measuring stick for success was not about where the baseline was; it was about how much improvement could be achieved from that starting point. So a number like 32% would be viewed as so deliciously low, that it made dramatic improvement almost inevitable.
Once the productivity story was told and understood, we were ready to jump into the trenches and start digging. But to get the team to jump in with us, we had to get them to own it. Up next, the challenge of creating team ownership.
Copyright © Tim Sullivan 2012