Professor Daniel Jones, founding chairman of the Lean Enterprise Academy, entrepreneur Jean-Claude Bihr, and author and speaker Michael Ballé collaborate to discuss the plan-do-check-act from a scientific angle and the importance of blending theory and investing in test methods to produce new kinds of lean innovation.
Lean was explicitly built on the systematic intention to distinguish what we know from what we think and to constantly explore the boundaries of what we know.
W. Edwards Deming’s plan-do-check-act discovery cycle is at the very core of lean thinking, as are the famous 5 “why?”. Plan and do come naturally to any business leader. After all, the best way to react to constantly shifting market conditions is to figure out overall goals, break them down into concrete sub-goals and get people to act on these. In other words the “I decide, you do” approach that hopefully gets cascaded down the organisation.
Check and act are invariably fuzzier. In many business situations it’s relatively easy to check whether people have “done”, but the impact of their action tends to be often complex, ambiguous and hard to measure. Politically, it’s often expedient to commit to an action even when the feedback is unclear, just as it can be dangerous to admit to a mistake. Check and act, drawing the right conclusions from one’s actions, are often relegated to the “nice to have” or “when we have time” folders. Yet, check and act are precisely what distinguishes lean thinking from run-of-the-mill management thought.
In scientific thinking, check and act come first. Much of modern science is about constructing the measurement apparatus (think of the hadron collider to confirm the likelihood of the existence of the Higgs boson), confirming the state of current knowledge and thinking up questions to explore and investigate. Science indeed lies at the border between the known and the unknown. Act means adopt, adjust or abandon. Scientific thinking is the constant checking of the knowledge we adopt, the knowledge we need to adjust and the notions we have to abandon (no matter how familiar they have become).
Check and act is a large part of the lesson we have learned from Toyota over the years. For example, take a problem with the quality of injected plastic parts. Have you checked the consistency of the material you’re using? Or a machine producing defective parts. Have you checked whether the ground it’s put on is perfectly level? A batch of parts have a few defectives in it? Do you know exactly how your process can produce defectives?
Toyota has never accepted the idea that a certain amount of non-quality is unavoidable in any manufacturing process and the cost of seeking zero defect is marginally too high. What they have taught us is producing bad parts is a reflection of a manufacturing process that is not fully understood. And this is precisely the barriers we need to explore.
As the CEO of Alliance, a high tech company, one of the authors has to regularly grapple with truly confounding defects, particularly when parts are made with radically innovative processes such as ceramics or exotic metallurgy. Traditionally, problems are sorted out by trial and error: think of something, try it, and see if it works and if it doesn’t, try again. Essentially, this is plan-do, plan-do, plan-do, until defects get down to an acceptable rate and production can continue.
For instance, one high-end jeweller requested ceramic parts for a large piece of jewellery to counter the mounting competition by innovative designs and materials. Being familiar with metalwork and precious stones, the client would assimilate any visible blemish as sign of a structural defect in the part (even non-visible parts). The company spent months and hundreds of thousands of euros creating different manufacturing techniques by trial and error and ended up systematically being rejected by the client. Tempers ran high and the project was in jeopardy until the author changed strategy. He then devised with the client what exact tests the parts had to succeed, and invested in specific equipment to test parts one by one (rather than full batches as they’d previously done). It emerged out of this discussion that the client had never faced such a problem and simply didn’t know – which created an opportunity to work together and patch up the relationship.
As they worked on it together they discovered not all surface blemishes mattered for the final jewel but, conversely, something no one expects, some visually perfect pieces could hide structural flaws and be too fragile to satisfy customers purchasing such high end jewellery. In the end, the focus on check first enabled the company to deliver and increase its standing in the eyes of its client, which led to further demand – while also deepening knowledge of its new technology.
As lean is implemented in the company and as the shop floor is run through a pull system, every single defect becomes the grain of sand that stops work from flowing – there are no “acceptable rates” of defectives any longer. Quality issues must be solved.
The answer lies in check. Beyond fixing immediate production problems, mastery of check is a true source of innovation. As test methods are better understood, the company opens new doors of possibilities with its technology.
For instance, the firm was contacted by a large aeronautics company to investigate the possibility of replacing conventional assembled metal parts with metal powder technology. The incentive is the company’s technology being additive (you bake the part out of metal powder) rather than subtractive (you carve the part out of a metal block), it is both lighter, cheaper and far more environmentally friendly. But, again, as with every technology substitution situation, agreeing on what works and what doesn’t is a key roadblock. With traditional metal parts assembly, the aeronautics company relies on complex and costly destructive tests.
Building on its previous experience with the jeweller, the author addressed the check issue head on with the aeronautics engineers – rather than start with a commercial discussion, they delved right into what kind of tests and test methods would satisfy product requirements. This approach radically changed the nature of the discussion as both parties, again, explored what they knew and didn’t know about the parts themselves, their life in use and so forth. Focussing on check ahead of plan and do radically reduced the cost of development, as well as create a sound basis for the relationship.
Treating lean as a science – the blend of theory, tinkering in the garage and investing in test methods, opens the way for a new kind of innovation, one based on small continuous steps of kaizen rather than costly breakthrough. Alliance’s technology is cleaner, lighter and cheaper. Yet it is also unfamiliar and technically demanding. Focussing on check and act, before plan and do, radically reduces the burden of trial and error, both in terms of development lead-time, cost and relationship tensions. But beyond its immediate benefits, Alliance’s experiment with test methods sheds a new light on another part of the Toyota puzzle: how come it came up with the Prius technology for mass markets and why do its competitors find it so hard to replicate its hybrid mastery?
Alliance’s CEO committed to lean as his main strategy several years ago and we have the unique opportunity to see how lean processes, both in manufacturing and engineering, affect product development and the product itself. A product is the result of a sequence of decisions taken by customers, product engineers, manufacturing engineers and production, and how well they work together.
By focussing on one-piece-flow, don’t make defects, don’t accept defects, don’t pass on defects (no setting problems aside to move the project forward and work on them later), lean processes affect the sequence of decisions and indeed the resulting project. This, we believe is the key to use technical innovation to continue to develop greener, better, cheaper products – in other words, increase value whilst minimising waste. In this sense, we believe the lean development process is unique and specific enough to merit the label of “leannovation”.