Q&A with Jeffrey Liker Part II: How to achieve Lean leadership and culture and the role of AI

Continuous improvement

Q: What is the impact of the new ways of reducing hierarchy to be more agile on Leaders?

The idea of the lack of hierarchy is not something new. The agile leadership model was promoted by Lee Iaccoca who set up the Iaccoca Institute at Lehigh University in 1988.  When I was in college back in the 1970s, there was a big movement of self-managing work teams. Volvo, for example, is very famous for having about three levels of hierarchy and they organized a plant around shops, each shop would make a part of the car like an engine or body or paint and they would be given targets for the day. They would be told make this number of cars and you figure out how to do it. That particular plant went out of business after about a year or two because the product wasn't selling. It wasn't because of the organization approach but there have been studies that showed that Toyota's system that has hierarchy is more effective and how the Toyota structure was far more productive than this plant. 

Since the self-managing team movement, many companies then tried to go without hierarchy. One of them is Scania Truck in Sweden, the truck maker, and they were very proud of the fact that they had flat management. A supervisor might have 50 people reporting to them and then they started learning the Toyota Production System and they started to see that a key to Toyota’s success is that they had a lot of layers of leadership and these leaders were accountable.  They were leading Improvement and had a vision for what the process should look like and we're leading the group toward that vision and they're continually improving, and they were much better trained in problem-solving than the normal production worker. Toyota’s ideal structure is that every leader only supports four, five people. They have a coaching role, they're building the team, they're developing those people. So, when you have a leader for each four or five people, you end up with a taller hierarchy. As Scania progressed with their version of the Toyota Production System, they saw the need for the team leader role and a smaller span of control.  They adopted this, trained the team leaders in problem-solving, and all their key performance indicators went up.  They have not looked back. 

Personally, I don't believe it's a good idea to have a structure without a hierarchy, often popular in software development. And even then, there's a lot of structure needed to make that kind of system work and there's a lot of strong leadership needed. So, I think the notion that you let people do what they want and they'll do the right thing, I don't think that’s very realistic if you need to get high levels of productivity. If you're depending upon producing something or a service that's repetitive then you care about efficiency, you care about on-time delivery and quality and you care about safety. I believe the best thing to do in that context is to have leaders who act as coaches and facilitators and are very effective leaders. In fact, you could say the ultimate agile organization is a culture of continuous improvement.

Q: What are the limitations you see to a continuous feedback system—emotional and technical?

If you look at systems theory, generally the faster the feedback the better. There's a rule for example if you want to punish somebody called Hot Stove Rule. When you were a child you might have touched a hot stove and you got instant feedback: this hurts and your tendency is not to want to touch that stove again. It's also impersonal. It's not your mother yelling at you. It's something that comes from the environment, it’s not useful to get mad at the stove. The key characteristics: it's impersonal, the feedback is rapid and severe for punishment. In case of rewards, the feedback should include some personal touch and should also be immediate, as close to the behaviour as possible and the longer the gap between the behaviour and the feedback the less effective the feedback will be. The problem is that humans can only respond to a very limited amount of feedback at a given time. 

Behaviour is very complicated and even within a few seconds a lot is going through my mind and I'm doing a lot. If you try to give me a big batch of feedback amongst everything, I'm doing it's basically worthless. You have to be very selective and intuitive in what feedback you're going to give me and when. 

In Toyota, for example, they have a process that you might have heard of called hoshin kanri or Policy Deployment where they identify the critical few items that they're going to have a special focus on this year. They will always focus on cost, quality, productivity and safety, and morale but they will have a few areas that they are specifically focusing on. They have a general visual board for KPIs, but they may have another board that has a special focus on quality and they'll pull out key information on whether your process is causing defects and you'll be able to see that as the day goes on. Being selective and prioritizing is critical.

The problem that we often see these days with computers is that it's relatively easy to give lots and lots of feedback. Maybe there are 15 KPIs and in real-time I could see how I'm doing on 15 things. But seeing how I'm doing on 15 things in real-time might be no better than getting no feedback at all because I just can't focus. I can't process that much information in real-time. One quick example that I wrote about in the Toyota Way to lean leadership was about a former Toyota executive who took a job as COO of a parts supplier.  He observed in one plant that they loved computers and they loved having nice displays by every process in the shop floor so you can see in real-time how you're doing on many indicators. And it was very clear to him that there was too much information. Everybody had heard through the grapevine that Toyota guys don't like computers, which is not really true. But that was their impression. They were afraid that he was going to take the computers away. He came in and he said I'm not taking computers away. In fact, I'm going to buy you the best displays on the market. But I'm also going to put a whiteboard next to the computer with markers. And what I want you to do is to pick two or three key indicators that you're focusing on every day. Pick the most important two or three indicators for your process and I want you to write on a whiteboard whether they're green with an arrow going up or red with an arrow going down or yellow where you are not making progress. When I come to visit, I'm only going to look at what's on the whiteboard. I'm not going to look at the computer, only the KPIs on the whiteboard, whether they’re going up or down and how you are responding.  After that, the KPIs, which had been stagnant, increased sharply.  The combination of focus, and personal attention when he visited motivated people.

Q: Are there any examples of industries other than automotive where TPS doesn't work so well?

There are many examples where leaders try to force-fit particular Toyota solutions that do not work well in their environment.  There are a lot of differences across different Industries in what lean looks like. I'm working now with a steel company and it's big machines and you can't make a one-piece flow little cell and it's naturally a batch process. You can't really do the kinds of things you often do with lean, starting with value stream mapping and then you create a cell and you create a pull system, those sorts of changes. On a big-picture level, it may eventually make sense, but it might require rebuilding the whole plant. 

One of the areas that we focus a lot on is standard work which looks at how people monitor the equipment, how they respond to the equipment, how they do changeovers and how they clean the equipment. Those are all critical to the high performance of the equipment. They might be able to influence 30 to 40 percent of the throughput based on how they respond to the automation, how they maintain and how they operate the equipment. So there's a lot of discretion involved and poor decisions can lead to various problems, including quality problems, that then lead to a lot of manual rework and that can be dramatically improved. They also have test operations, which in some cases becomes a bottleneck, and that's a fairly manual operation. Standard work, visual management, 5S, teaching problem solving are the most critical things that we’re working on. I've worked in Iron ore mining that's outdoors, not going into a mine. They were blasting and then trucks picked up the pieces and they were sieving out the different size pieces through machines that crush and so on-going on for miles.

I've also worked in Iron ore mining that's outdoors, not going into a mine. They were blasting and then trucks picked up the pieces and they were sieving out the different size pieces through machines that crush and so on-going on for miles.

And again, we did not shift to small teams of people working in a cell, but we focused on equipment maintenance and how to figure out where to blast at the right time and cleaning the equipment that sifts through the material. There was a lot of work that had to be planned on a day to day basis and they built a visual planning center and held daily meetings and all these changes made a huge difference in how the mine operated dramatically increasing throughput.  So, you have to be sophisticated enough to understand the principles of lean and apply them in an effective way to this process and not simply try to copy Toyota and make it look like their plant. (This case is described in detail in The Toyota Way to Continuous Improvement.)

Q: How can a leader find a balance when he or she has a shared position between two countries when one has hierarchical culture and the other has zero hierarchal culture?

What you want is to put them in a blender and mix and match (laughs).  One of the principles that Toyota has is respect for people and under respect for people, one of the principles is to honour the customs and norms of each country where they do business, Toyota has plants in many countries with very different cultures such as Russia, Indonesia, England, France the United States, and India. There are many operations throughout the world, and they believe that they need to adapt to that culture. At the same time, they believe that there's a core philosophy and a core way of thinking that they call the Toyota way and the Toyota Production System that's not negotiable. So, the concept of one-piece flow is not negotiable, the concept of continuous improvement and learning problem solving, the concept of respect for people and investing in developing people are not negotiable. The idea that management should be leaders who are creating a safe and secure environment and then within that developing people to achieve very challenging goals, that's not negotiable. But there are specific human resource practices that are negotiable in some countries and they need to adapt to that culture. They worked hard to identify the critical features that are not negotiable, and they've managed to teach the Toyota Production System and make it part of the culture in every place they operate, even though it will be more or less hierarchical in different places.  Interestingly Japanese culture is very hierarchical.

Some places require more seasoned Toyota leaders, say from Japan, running the plant until the plant within the country becomes more self-sustaining. They also have what they call a mother plant system where every plant has a more mature plant that is its mother that has the responsibility to teach it TPS sending lots of leaders over, sometimes for years.  Originally the mother plants were always in Japan, but more recently local plants have developed enough to act as mothers.   In America, for example, the Georgetown Kentucky plant is the oldest plant and they could act as mother plant to another American plant and in Europe the UK plant is very mature and can act as a mother plant and they have a lot of activities where they send people across plants to spend time working and learning within that plant. It takes years, not weeks, and it probably is not the case that the country that's very hierarchical is going to look exactly like the country that's not hierarchical. You need to move the command and control plants closer to the middle, I think.

Q: How do you see lean leadership role changing in the face of automation and AI wave? Will leadership trickle down to process managers (like designing their own processes assisted by machines) or will it centralise more to top management removing intermediary leadership positions?

From the point of view of a high control-oriented executive, or a manager who feels that they have a great need for control, the ideal image is automation that works at the push of a button so he can sit at home in front of a computer, see the status of the operation and then take corrective action and maybe ask an engineer to use a computer to fix the problem.  The ideal image is to take all the people out of the system because people are the most uncertain part of the system. You never know what people are going to do. But a robot that is properly programmed as well as maintained will do what it's supposed to do. It won't talk back. It will not show up to work drunk. It will do repetitively what it's programmed to do. The problem is that, firstly, with today's technology the more complicated the robot is, the more likely it will breakdown and therefore somebody has to fix the robot. So far we don't have robots that can fix other robots. That requires people. Secondly, the robot has to be trained. It’s only going to do what it's programmed to do. 

In January 2019, we visited a Lexus plant as part of my “Leading the Toyota Way Japan” Masterclass and they were showing us the painting process which is all robots. And it has been almost all robots for about 35 years. It's not a new idea. Lexus believes in perfection and they had master painters who could reach every nook and cranny and distribute paint across the body perfectly. They wanted those master craftsmen teaching the robots with the idea that a master painter can program the robot to work at a high level like a master craftsman. Even who programs the technology can make a big difference, so people don't ever leave the system, there are just fewer people doing the repetitive tasks, but you still have to set up the process and it could be set up in batch mode or could be set up in one-piece flow. What Toyota's view is, first figure out how to do the process manually, then when you work out the bugs in the process, then bring in the computers and then have very knowledgeable people who are maintaining, programming and operating the computer system who are there on the ground and could respond immediately to problems. The IT ideal kind of vision where the elite, the manager is sitting in front of a computer and there's nobody in the plant and you don't need lighting because robots don't care about lighting. That hasn't been proven to work on a widespread basis so far. If you are familiar with Kurt Vonnegut’s book, Player Piano, you have an idea of how bad the world could be if computers ran all our workplaces.

Q: Could you give some advice on the management application of Plan-Do-Check-Act with shop floor team members?  

You can think of Plan-Do-Check-Act (PDCA) in two ways, one in a top-down command and control environment, where you tend to look at it mechanistically like it's a tool that you plug in, or you can look at it, as Toyota does, as a way of thinking. For Toyota, PDCA is the essence of what they call problem-solving and it's a mindset of continuous learning, continuous testing, continuous experimenting.  Let's say that I was asked to lead a team to reduce the cost of making batteries in half, which is a big priority right now because batteries are the most expensive part of electric cars.. My natural tendency at Toyota is going to be to study and deeply understand battery technology and whatever better technology there is in the world, then come up with ideas for improving the process of making the batteries or redesigning the batteries and then run a lot of experiments, trying a lot of things, and learning from each experiment. The faster you can run the tests, the faster you get feedback, the faster you can learn. There is a big emphasis on doing, not just thinking of ideas. If I can find a cheap way to test my idea that's very encouraged. Every time I test an idea I’m doing PDCA. I have a clear hypothesis, I run the experiment, then check to see what happens and then I ask myself: What did I learn from that and based on what I've learned, what would be the next experiment to run? So PDCA could be happening multiple times in a day and certainly should be happening every day, at some level. It really becomes part of the culture.

As mentioned by Mike Rother in his Toyota Kata approach, what he's trying to teach is scientific thinking, which is based on rapid PDCA.  He argues that PDCA thinking is not natural and needs to be taught through practice, ideally daily, with a coach.  It is only through repetition that we can rewire our brains to approach our goals scientifically.  The only practical way to do this is if every manager, supervisor, team leader learns PDCA and learns to coach it, which takes many years.

Q: How do you suggest conducting experiments when we need to improve some difficult practice? 

I recommend involving the people who are doing that practice. It's one thing to work out offline a best practice and then teach it to those people. It's another thing to get them to figure out the best practice and learn from the experience of trying to run experiments and learning some things work better than others. Part of the problem with complex processes is that usually a small number of people have native knowledge. They figured out how to make it work, but they don't really even know what they do, let alone how to teach other people and they may not be very good at learning from other people because they think they know best. The goal is to work together on developing the current best practice we call standardized work. It’s always better to work with those people involved in the practice, not to just ask them to come up with the best way. They also need guidance. They need leadership to figure out a good process for doing PDCA and learning the best way they can today, and then continue improving on it.  Ideally whoever is responsible for that process is also being taught scientific thinking so they can continuously Improve.

Q: How to address organizational silos?

I mentioned the process of hoshin kanri. Toyota has an annual cycle where for three months they're developing their targets for the year and they develop an initial plan for how to achieve the targets. The plan is not solutions, but it's areas they're going to focus on. Some of those problems are cross-functional and some of those problems are within a function. Obviously, the ones within a function that has hierarchy are easier to manage than cross-functional ones. However, one of the things that Toyota believes very strongly in is horizontal leadership, which is learning to lead not only people below you but also your peers or maybe even those above you. In that hoshin kanri process, Toyota will assign somebody to be the lead. 

For example, as head of manufacturing for North America for Toyota my co-author, Gary Convis was given the challenge to reduce warranty costs by 60 percent over a six-year period and warranty costs come about mainly because of design issues. They come from the engineers who work in America in the Toyota Technical Center. They also come from production engineers who are working in Japan and America, and a lot of them come from suppliers.

Gary was the head of North American manufacturing, so for him to do his job effectively, he had to pull together a team of the head of R&D in the United States, the head of manufacturing in Canada, the head of quality, purchasing heads, and some key leaders in Japan and he had to lead that team. That was horizontal leadership and he had to break down the barriers of the functional silos. He was at a high enough level that he knew all these people and they had great respect for him. So, they weren't necessarily breaking down rigid silos, they were helping Gary achieve his challenge which came from the very top of the company so they knew it was very important and they needed to bring their organisation to support him. It looks more like a matrix organization where there is one person who's responsible and other people might not have a direct reporting relationship to that person, it may be more like a dotted line reporting relationship. In part, they gave Gary that assignment to further his development. They were saying, you’ve achieved this very high-level position because you have great leadership within the manufacturing function. We need to further develop your skills at leading horizontally, which is the true test of leadership. Leading horizontally is maybe the most critical part of breaking down the silos. You can do various things by rewriting the organizational chart, but ultimately it comes down to people leading across organizations. 

Q: If Toyota wouldn’t exist at all, which other systems you know would be a good benchmark for management?

There's a lot of good systems. If you go to one of Jim Collins books where he talks about great companies, they all are good at something. Walmart had their great leader, McDonald's had their great leader, so did Amazon and Apple. According to Collins leaders of Great Companies have some common characteristics including they know their business, not business in general, they are humble, they put the company before their own egos, and building the company is their life’s work.

The challenge is that, when you have a start-up like an Apple or Google or an Amazon, and then a leader like Jeff Bezos of Amazon, he gets to a certain point when the company gets so large that it becomes difficult for him to influence the culture among hundreds of thousands of people in so many countries. He's delegating to a lot of other people major responsibilities and they may or may not have his vision for leadership and skills.

One thing that has been unique about Toyota is that that they have had the same lineage of leadership going back to the founder of the company over a hundred years ago and the principles of the Toyota Way can be traced back to principles of Sakichi Toyoda when he started the original loom company. Their philosophy grows from within, so the current Vice President probably hired into Toyota working in production or working as a junior engineer and they spent 30- 35 years with Toyota. That has allowed Toyota to continue to reproduce their culture and maintain a very clear focus and very clear set of values and rules of how they operate and work. That's interesting about Toyota. If you try to copy Toyota, at a point in time you’re trying to speed up the process that's taken a hundred years. In that sense, maybe there's some good in picking your favourite company, whether it's Amazon or Google, or maybe there's a smaller company. There are always companies that are considered best places to work, they might only have a few hundred people working for them, and they may be an interesting model for you. There are lots of great companies, small, medium and large that have demonstrated the power of having a good philosophy, a good model of leadership and sustaining that over time and growing from within, as Jim Collins has identified. His concluding characteristic of great companies is that leaders are building a company that can last and they're developing the leaders who then succeed them. 

Hope you found the answers useful. For more information on building a sustainable culture of continuous improvement and how to become a Lean leader, attend Jeff’s “Leading the Toyota Way” Masterclass.

Article Share Image Share this article on your TLN® Profile
Share this article TLN Linkedin Post TLN Twitter Post TLN Facebook Post