Home > The Lean Post> Kanban As A Learning Strategy
The Lean Post
Sharing how the world is making things better through lean.

Kanban As A Learning Strategy

by Michael Ballé, Jacques Chaize, Orest (Orry) Fiume & Daniel T. Jones
March 22, 2019

Kanban As A Learning Strategy

by Michael Ballé, Jacques Chaize, Orest (Orry) Fiume & Daniel T. Jones
March 22, 2019 | Comments (5)

Just as the world that businesses compete in has changed dramatically, so too must the way we conceive of what we call strategy. Strategy today is routinely assumed to consist of setting goals, developing a plan to meet these goals tracking towards the progress of these goals and re-planning accordingly. It conforms to the competitiveness model popularized by Michael Porter in which companies dominate by identifying market opportunities and locking in their position against competitors. Yet the power of this approach when faced with dynamic markets and what might be seen as “wicked problems” remains an ongoing debate, one that has raged amongst strategy experts since Von Clausewitz’s landmark writings on “friction” and “the fog of war” that show that no plan can ever sustain contact with the enemy – or reality for that matter.

The dazzling success of lean companies (Toyota being the exemplar) provides a powerful alternative to the notion of strategy consisting of having the best plans and greatest bargaining leverage. Toyota’s success might best be attributed to its lean strategy, a dynamic, agile, and learning-based approach that can be summarized as: understand your challenges, your core practical problems, train your officers to do so and know how to respond, trust their initiatives and then reflect on what happened and learn fast. Traditional strategy is about optimizing the situation as we see it (assuming perfect knowledge and the assumption that nothing will change quickly); the other is centered on creating a dynamic of trained, autonomous people that can align spontaneously on a shared intent. Facing your challenges, responding faster with all your people, studying your response and learning to do things in new ways together is the lean strategy needed to face the disruptions of the twenty-first century.

Toyota engineers realized early on that the kind of issues that propped up from the use of Kanban were hard to solve with traditional engineering solutions – they were too granular, too small. They went about solving this through kaizen: continuous small steps improvements from the delivery teams themselves, supported by engineering when needed.  Most of the twentieth century’s leading companies based their strategy on the premise of optimization through standardization: the key to making money was to design products for the widest possible segment, find the lowest cost production location for each component, and batch work to get unit costs down. These companies leveraged economies of scale– the massive cost advantage from doing the same thing in the same way all the time, which has driven the success of the McDonalds and Coca-Colas as long as every customer is happy to purchase the same Big Mac all around the world, all the time. This one-size-fits-all “best practice” approach had benefits yet turns out to be incredibly inflexible and wasteful of both human initiative and capital.

So far, the twenty-first century’s disruptions are all about treating each customer as a unique individual, striving for near-instant responses and connectivity – a very different world indeed. Advances in computing power have allowed the Google, Amazon, Apple, and Facebook disrupters to personalize their answer to each person. Google gives you answers to any question you will have, Amazon ships to you any product you want from an infinite shelf, Apple lets you customize its devices in any whimsical way you like through apps, and Facebook allows you to exist in virtual space through the avatar you design. Products and services can now be both scalable and custom-made.

Yet while the disruptive few have thrived with this approach, the majority of established companies have failed to grasp the underlying way of thinking that drives this model. They adapt to the new strategic model by rearranging processes and systems without rethinking and transforming them. Many companies still believe that digitization means running their one-process-fits-all, batch systems on computers. Undaunted by the repeated failures and increasing complexity of traditional IT solutions, they continue their search for “best practices” to the point that many management teams are reduced to upgrading systems, hiring subcontractors and fire-fighting. To any immediate problem the answer is “we’ll have the perfect solution in 2 years’ time” but no one knows what to do right now to solve a customer issue.

These companies fail to grasp the true function of the new tools, which largely transfer the authority they used to outsource to outsiders back to their own hands. Consider the new way we think of photographs.Taking a photo with your smartphone instantly breaks down the image into unitary pixels, enabling you to share it immediately from the phone on any social media you like. This is worlds apart from taking a picture with a film camera, having to have it developed, and then keeping paper copies in an album.

We humans fail to use tools properly if we don’t grasp the idea behind it, or if our mental model is no longer current. For instance, machines in factories in the days of steam were organized in straight lines, with each machine powered by a transmission belt running from a central rotating shaft. Electricity then enabled machines to be placed anywhere to create flow lines, but many factories are still designed as batteries of same-process machines as if there was a central line shaft. Likewise, the mental model of the line shaft lingered long after the technical constraints were lifted. Today many factories still operate as if this constraint were necessary: they continue to create process villages with batteries of same machines producing heaps of parts. They still design organizations around functional departments, with the belief that if each department did its job efficiently, the whole would somehow be effective.

Such “phantom limbs” of production continue to constrain companies. A large part of management literature of the twentieth century has been about how to fix the previous mess, from process reengineering, IT workflow controls, flat organizations without much success. Today functional departments are mostly run by IT systems (the real boss of any plant is its MRP system), more silo-ed than ever and more incapable than ever to react quickly to spot customer demands through teamwork. Nobody remembers the line shaft, but the moment you try to create a multi-process flow cell to flow all value directly to customer, you hit all the barriers of functions trying to maintain ownership and win turf wars—rather than cooperate to flow value to customers. Similarly, the flow line is a result of electricity, and the technical ability to create small machines on each process step, place them where they make more sense to build the product in sequence one by one and ship it right away to customers.

What is needed is a fresh start along with a new mindset. The thinking we see in the digital-native firms that are disrupting the twentieth century business model is radically different. Economies of scale no longer make much sense; economy of data analysis and learning does, spectacularly. It’s about winner-takes all through customer-centricity, velocity of innovation, concentration of talent, and financing growth through access to cheap capital rather than painstaking generation of cumulated profit. Traditional firms are trying en masseto get with the program, but it’s simply hard to get there from here. So how can it be done? 

Kanban as Strategy

With Kanban, problems appear at a very detailed and granular level: one card cannot be served, or cards accumulate anomalously at some point in the flow of cards. They reveal problems as they arise, where they arise. The people doing the work themselves can—indeed must—learn to take responsibility for the problems and solve them, which then creates a flow of ideas throughout the company. Back in the 1960s, a bunch of engineers at Toyota, which was then a little-known Japanese automaker scrapping by after recovering from bankruptcy, had a breakthrough insight. They would replace lists of things to do, production programs, etc. with Kanban cards. A Kanban card is a signal with which the customer process goes to pick up parts (components, information, jobs) at the preceding supplier process where the work is held in a supermarket. The supermarket is replenished by the supplying process on the basis of “a little of everything all the time” so that the needed work is always available in small quantities. It operates in the “real time” of the work flow. As a result, real customer information is transmitted through the system and work flows back in response. One card being the demand for the production of one box of parts, much as in a restaurant, one slip of paper is one order for one meal.

But Toyota’s aim was not only about changing the work instruction and scheduling method, as the West has generally interpreted it (projecting its old thinking on a new insight): they also immediately started to measure the response time to one card. They tracked the lead-time between the specific demand for one box of products and when this box was physically delivered.

This humble looking Kanban system turned out to be a complete revolution because it opened a new path to improvement. In mainstream management thinking, improvement stems out of “design thinking” – smart people figure out the entire system, find out what’s wrong, and make changes for a better system.

With Kanban, problems appear at a very detailed and granular level: one card cannot be served, or cards accumulate anomalously at some point in the flow of cards. They reveal problems as they arise, where they arise. The people doing the work themselves can—indeed must—learn to take responsibility for the problems and solve them, which then creates a flow of ideas throughout the company, and very different ideas as that. Rather than “we must invest to change the whole system” engineering thinking, people at the coal face tend to find astute, capital free ways of making things work, and opening creative solutions no one would have thought of by redesigning the system.

Toyota engineers came to focus on two types of problems:

  • Distinguishing normal from abnormal conditions:Items that lagged on the standard delivery time, which meant focusing immediately on issues. Imagine that you need to deliver a job in two days. If something happens and work is not sent to your customer on day three, you know immediately that you have a problem here and now and call for help to get back into a normal condition.
  • Bringing value closer to customers:the standard response time can be constantly challenged and reduced, to surface new technical problems. If we normally deliver the job in two days, let us set the challenge of delivering it now in one. This will seem impossible, but lead us to scratch our heads and find new ways to do old things – and occasionally stumbled on innovative insights.

Toyota engineers realized early on that the kind of issues that propped up from the use of Kanban were hard to solve with traditional engineering solutions – they were too granular, too small. They went about solving this through kaizen: continuous small steps improvements from the delivery teams themselves, supported by engineering when needed.  Getting engineers to support line operators in their kaizen efforts taught engineers many new, unexpected and clever ways of solving engineering problems. Kanban was a key tool for kaizen, as they discovered. 

Again, the physical tool disproportionately influences the thinking, and over time Toyota leaders began to think about their car lineup with the same mentality that launched the Kanban card. Rather than design new models to capture the entire market, they saw each model as a specific response to customer needs they needed to upgrade at a regular rhythm (and trying to reduce this as far as they could) to continue to serve the customers with the same needs. The Corolla, for instance, was introduced in 1966 and is currently in its 11thgeneration with the same principle of an affordable family car with some luxury features, which has been reinterpreted every 4 years for half a century. In a sense, there is a Kanban on the corolla design with a 4-year lead-time. This is the model that Steve Jobs embraced with the iPhone, releasing a new model every year, almost to the day.

We had the opportunity to see the impact of Kanban firsthand in the early 1990s as we studied how Toyota started their operations in Europe and drew their local suppliers in their just-in-time supply chain by begging them “not to mess with the Kanbans.” Previously, the supplier would receive a global order, schedule its production in large batches, and ship to the customer when the stuff was ready. Working with Toyota, suppliers had to learn to deliver to Toyota trucks that came several times a day with precise manifests requiring a little of everything in each truck, according to Kanban cards (paper slips by then printed from electronic communication in the morning). This forced the supplier to organize the consolidation of truck preparation according to the Kanban cards, reduce the batches and deliver boxes one by one. Like putting on a suit two sizes too small, this created friction everywhere and led the supplier to solve a variety of problems:

  1. Quality issues now appeared starkly (no Kanban card would be served with bad parts). Rather than produce an entire batch, inspect it, isolate the doubtful pieces and deliver the rest (in which case some bad parts always slip through), the supplier lines now needed to deliver to shipping only good parts—every part needed to pass quality standards, every time). Reworked pieces created endless problems, because it would take time from making a good part right-first-time, and so increase the response time, and so on.
  2. Flexibility issues became unavoidable. In order to produce a little of everything all the time, changeovers had to be made so easy the operators themselves could do them in response to the Kanban sequence. Layout of cells was changed again and again, processes improved, until work could flow more or less continuously, shifting easily from one product to the next in the product family. 
  1. Involving people became as essential as it had been in Japan. At first European suppliers tried to crack the problem the usual way with engineers devising engineering solution, but they soon found out they had to work with the frontline teams to solve all sorts of very small issues, such as screws not screwing right, machines not being 100% reliable, cross training, inconvenient logistics and so on.

Kanban As a Learning System

In many ways the underlying model for Kanban cards—to use representations of work as a means of controlling its progress and providing real-time information among the workers—was a precursor to today’s digital economy. Beyond the amazing productivity gains Toyota was generating in its entire supply chain, the true impact was felt in how the company worked as a system. For example, product engineers got involved, and with the steady rhythm of model change and product renewal, would start to use knowledge learned from kaizen efforts to design products with greater functionality and lower costs. Toyota’s “just-in-time” supply chain was far more than simply controlling the logistics flow (which already delivered massive productivity increases) – it was about organizing a steady flow of value towards making better cars and better satisfying customers. The lesson here was: there is no best, only better.

In many ways the underlying model for Kanban cards—to use representations of work as a means of controlling its progress and providing real-time information among the workers—was a precursor to today’s digital economy. The Kanban driven just-in-time value chain is: pixelized (lead-time is followed box by box), real time (the challenge is always to reduce the lead-time between demand and delivery to get to instantaneous) and connected as Kanbans connect the entire value network to better satisfy customers. This dynamic was the foundation for Toyota’s main strategy of “one time customer, lifelong customer” (focus on satisfying existing customers rather than seek to escape problems by convincing new ones – “churn” in modern terms), realized through “sell one, make one”, which could only happen by using Kanban and supporting, encouraging, cajoling all people all the time into kaizen efforts. Herein lies Toyota’s truly disruptive innovation.

Toyota’s Kanban legacy—it’s underlying ideas—have far more direct lineage with today’s digital economy than most folks realize; and capture the core elements of the disruptive lean strategy fueling many of today’s successes. And yet we are amazed at the number of “lean” books that reassure readers they can do “lean” without trying their hands with kanbans, although all the spectacular success stories we’ve had the privilege to see firsthand started with kanbans. Similarly, it’s not surprising that disruptive companies such as Amazon deliberately built their supply chains on Toyota lines, hiring executives well versed in lean thinking to do so. As the dust settles, the tool is essential. Electricity allowed Henry Ford to create the conveyor driven flow line. Toyota engineers picked that up right away in the 1930s when they were building looms. Airplane manufacturers only started experimenting with the first flow lines at the end of the previous century, and, regardless of the advances this triggered, this has yet to spread far in the industry.

Toyota has offered the world a method to get from the traditional industrial mindset to the modern thinking of digital disruption. They’ve developed it ahead of everyone else, and have strenuously written it down for the practical purposes of educating their suppliers and transplant operations. We discovered this in the late eighties, when we coined the term “lean” to describe these fundamental ideas, and have been studying the various efforts to translate this thinking outside of Toyota and more widely outside of the auto industry. What we found confirmed that mental models are incredibly sticky—that even the smartest people find it hard to change their minds. In reviewing the past twenty years of lean experiments, we have been asking ourselves what ideas need to change in order to adopt the lean thinking method to transform companies and prepare them to the challenges ahead. We’ve come up with four critical thinking changes:

  1. Reusable learning rather than best practices adoption
  2. Performance results from dynamic engagement rather than static optimization
  3. Innovation from people-centric, leadership from the ground up
  4. A radical change in how we understand executive decision making, with the Find-Face-Frame-Form decision cycle rather than the usual Define-Decide-Drive-Deal, and a shift in the role of hierarchy from a chain of command to a chain of help.

The first reason we’ve found people refuse to look into the telescope – specifically, adopt kanbans – is that they see the Kanban system as a “best practice” that will replace their current IT driven scheduling with paper cards. There are two deep misunderstandings here.

There is No Best Practice

First, the very idea of “best practice” is a surviving notion from Frederick Taylor’s “one best way.” Of course, there is no best practice – every context is different, and things change too fast anyhow. The obsession with “reusable learning” is similar to the one with economies of scale. It hinges on the incorrect assumption that learning is about applying a static piece of knowledge to new situations. Today more than ever, learning is about transferring knowledge to new contexts, which means blending an initial grasp of the idea and reinventing it every time to better understand its deeper sense. What we need are ways to learn faster. Reusable learning is not about applying a “best practice” to all possible cases, but about looking for a smarter way to do things more effectively. To do so, we need 1) a starting point, a key to enter the room, and 2) a direction in which to look for solutions (and to avoid repeating known mistakes).

Today more than ever, learning is about transferring knowledge to new contexts, which means blending an initial grasp of the idea and reinventing it every time to better understand its deeper sense. What we need are ways to learn faster. The second misunderstanding is that Kanban is about the cards – Kanban is about measuring the response time on every individual demand, a very different concept. In essence, the flow of Kanban cards is a key circulatory system in a learning organization. Itemizing demand and tracking response time (there are many digital tools to do so now) is the entry point. Kanban is the start of learning, not the end. In using Kanban we’re not implementing a solution, we are looking for the next step to improve so that we will discover new ways of doing things through teamwork between all stakeholders. The Kanban card has to be served. The problems are immediate and urgent. This focuses everyone on solving innovatively clear and present issues rather than blue-sky invent applications for problems that will never appear. Improvement takes on a radically different meaning, starting from where we are right now. We need to manage learning curves, not action plans. We need to find smart ways to do the hard things (which are hard for every competitor) not work our way around them. We need to accelerate learning, not make obsolete processes and procedures fixed and rigid.

Better performance results from the dynamic of engaging with issues to better satisfy customers right now—rather than with many executive’s obsession to statically optimize through reorganization and footprint restructuring. Improve the work you have now by engaging all people in better seeing their line of sight to final users, better understanding how they help or hinder their immediate clients, and what can be improved in the way their work to make their jobs safer, easier and more interesting. Yet most companies would rather work with consultants to imagine the system upgrade that will solve all their problems in three years’ time (something, for that matter, that they will never know how to implement). Look, try, measure, think, look, try, measure, think, is about teaching scientific thinking to everyone all the time, and to create a dynamic of teams of teams rather than smother engagement and initiative with expert-driven processes, procedures and all the usual reasons not to try something new.

This new way of sustaining dynamic improvement requires a very different type of leadership. Daily kaizen is about learning to change, in small ways every day, by solving real-life problems for customers, for co-workers, for partners. It only works if leaders support this firsthand every day.

Yet this learning-based approach, paradoxically, often requires a humbler and less certain mindset from those in charge. Using response time as the entry key, the “cognitive handle” that enables people to tackle large complex problems pragmatically (as a real-life handle allows you to pick up a large and cumbersome suitcase) we saw that most of management decision theory rests on a “I am master of my fate” deep metaphor. We think quite naturally that, being the masters of our fate, our job is to figure out the right strategy and then get the organization of execute it. When things don’t work out the way we’d hoped, it’s not about that the strategic thinking was wrong, but that unknown (unknowable) factors and the unavoidable resistance to change led us astray. By looking into item by item demand and studying response time, we came to see there is a completely alternate deep metaphor in terms of “we’re challenged and we respond – with our allies.” In this worldview, we accept that we’ll always be challenged by a world that changes faster than organizations can, and that we don’t know the correct way to respond because of our own legacy mental models. The key to velocity then lies in facing the challenge rather than looking around the elephant in the room and then looking for new ideas with our allies in order to craft new responses rather than apply tired old “best practice” recipes that no longer work.

The traditional executive cycle can be described as (1) Define the problem from the information you have, (2) Decide on a strategy, (3) Drive this strategy through the rank and file to implement it and then (4) Deal with the consequences when things never work out as planned. By contrast, we offer a radically different vision of executive choices in (1) Find all “now!” problems (through just-in-time techniques) (2) Face the real challenges that emerge and set yourself a measure of these (3) Frame them in terms that anyone in the company, from executive committee member to janitor can understand (so they can contribute) and (4) Form the solutions by shaping organizational coordination around new tools and new ideas as we go, with employees and partners alike.

This new way of sustaining dynamic improvement requires a very different type of leadership. Daily kaizen is about learning to change, in small ways every day, by solving real-life problems for customers, for co-workers, for partners. It only works if leaders support this firsthand every day.Opportunities from new digital capabilities on the market now come far faster than most organizations can absorb. Typically, many organizations we see try to digitalize their current ways of working rather than embrace the revolutionary ideas underpinning digital tools. The deepest mental model shift needed to embrace the digital century is to abandon the notion that we are masters of our fate and that we will impose our brilliant strategies on employees, customers and partners alike but to think in terms of challenge and response, and studying the quality of our responses: facing our problems together and forming solutions with our allies. This leads to a radically new understanding of performance as the result of a people-centric dynamic in the business where everyone understands the plot (what the game is and see how they can contribute) and management steers the flow of ideas rather than impose rules and procedures. Such dynamism builds on constant learning-as-we-work so that we create true learning companies that manager learning curves at the same time as day-to-day work.

In the end, it’s all about the tool and if this dynamic vision is a lever, the pivot point is the simple idea that we should measure response time, not piece rate or unit costs. And the best way to fully grasp what a disruptive digital strategy can be is to, as many fast-growing IT companies now do, experiment with paper Kanban cards to understand intuitively what pixelisation, immediacy and connectivity really feel like in terms of responding to customer dissatisfactions, putting resources at the right place right now, and supporting individual learning and teamwork. To learn to disrupt, start by disrupting your thinking. To disrupt your thinking, disrupt your practice: try using Kanban on your own activities.

Kanban don’t lie. An action plan allows you to prioritize items, and choose to do the easy things first until you stumble on the hard things and nothing ever gets done. Line by line budget management allows you to spread gains and losses on convenient accounting lines. But Kanban cards mechanically reproduce customer demand, and demand to be served. The lean logic is the following: by responding to cards in the same sequence as they appear, the people who do the work themselves learn to better understand what their customers expect of them, better recognize problems, better interpret abnormal situations and react more astutely. As a result of doing this day in and day out, the inventories or backlog are lowered, and significant capacity is freed to make room for new products. Furthermore, the technical learning accumulated in the process of solving all the large or small problems revealed by the Kanban cards are invaluable to design better new products with more robust production processes. Not only does using Kanban free capacity for innovation, it also teaches how to innovate in practice, in order to better satisfy real customer needs.

In the end, this apparently mundane-but-hard change of practice will open your eyes to a radical rethinking of what strategy truly is: the ability to confront your clear-and-present challenges with trusted allies, to train your managers to understand better their own responses and the collective intent, and to learn at all levels, from the executive committee to the janitor to discover a more adaptive way to do the work so that the business, as a whole, responds faster and smarter. There is no such thing as “best,” there is only better, every day, everywhere, with everyone.

Search Posts:
September 11, 2019 | 3 Comments
May 31, 2019 | 2 Comments
April 3, 2019 | 7 Comments
Was this post... Click all that apply
HELPFUL
32 people say YES
INTERESTING
31 people say YES
INSPIRING
30 people say YES
ACCURATE
22 people say YES
Related Posts
5 Comments | Post a Comment
Mark Reich March 22, 2019
3 People AGREE with this comment

Thanks to all of you for this insightful and educational piece. Much more than a tool.  An endless refrain. Mark

Reply »

Bernd Sauerbrey March 22, 2019
1 Person AGREES with this comment

Thanks so much, Michael, for this advertsiment for kanban, I gave it to my young co-workers learning with quick serve now.

Reading your article I remember Eckhart Tolle's NOW where we also should live in current moment, in kanban we solve the problems at the point of happening and go next step, to the next "moment".

Bernd

Reply »

Jorge Wong March 22, 2019

Thanks for an enlightening and inspiring post. Your article really puts to rest the quintessential business dilemma of 'strategic choice':

The Tyranny of the “OR” vs the Genius of the “AND”.

The prevailing model on strategy making has been a trichotomy of strategic choices:  operational excellence, customer intimacy, and/or product leadership. To be a market leader, an organization had to be very good at one of these and acceptable in the other two choices.  Can you be very good at all of them? No, that was considered "impossible" or "impractical".    Enters Toyota with its TPS house, supported by JIT+ Jidoka pillars, and dwellers committed to Kaizen... and makes it possible to excel in ALL: customer intimacy, operational excellence, and product/service leadership.    “The test of a first-rate intelligence is the ability to hold two or three (apparently) opposed ideas in the mind at the same time, and still retain the ability to function.”
~F. Scott Fitzgerald   

 

Reply »

Linda Murphy March 23, 2019

Nice article and some great points that all of us in the Lean community would do well to learn. It should be noted, however, that the LEI book 'Toyota Kata' explained and illustrated this well about ten years ago. Clearly, knowledge is not the same as understanding. The latter takes practice, which is why "Kata" may be an appropriate word with regard to actually starting to address the issues you raise. BTW, an interesting (& entertaining) video along these lines is: https://www.youtube.com/watch?v=MFzDaBzBlL0

Reply »

art byrne March 28, 2019
1 Person AGREES with this comment

Gentlemen, great article, lots of good insghts. As someone who has always seen lean as strategic it is easy for me to agree with all of it. But, change is hard for most companies/people. All too often I have seen companies adopt kanban but not change anything else and think they are now lean. All they did was slap kanban cards on top of their big batches. As we progressed at Wiremold we got to the point where nothing could be built without a kanban signal. Kanban cards were generated off of the incoming orders throughout the day. As soon as the incoming orders for a particular product reached the level of one kanban worth [it could be a full pallet, a half pallet, ten boxes or whatever] a card printed on the shop floor in the cell that made that product and the product would be made within a short time afterward. In this way we connected the customer directly to the shop floor and drastically shortened the lead time [from 6+ weeks to 1-2 days]. At the same time, as you point out, problems in the kanban system gave us instant knowledge of where to focus our kaizen and daily management efforts. I hope those who read your post will take action based on it.

Reply »

Search Posts:
September 11, 2019 | 3 Comments
May 31, 2019 | 2 Comments
April 3, 2019 | 7 Comments
The Battle for the Soul of Lean
What Do Sensei Actually Do?
Lean Roundup: Pull
Lean Thinkers Are Systems Thinkers