How Come Toyota Veterans Don’t Use a Roadmap?
Dear Gemba Coach,
Recently I spent a few months working with Toyota veterans at an automotive supplier. I was surprised at their approach. I expected them to start with what I understood as the basic sequence outlined in books showing the TPS House: 5S, then stability in the 4M's, standardized work, kaizen and problem solving; after which we work on pull, jidoka, and more. The actual approach they took couldn't have been more different. They just went to the shop floor and started right in by tackling problems in very mundane, every day situations: defects, reworks, losses, line stops, and so forth. I realized that we were doing kaizen to put these improvements into place, and to implement continuous flow and pull between processes. Then we standardized it - though we were actually doing standardized work when we were creating continuous flow (including visual management for everything, 5S, measuring the indicators, doing constant follow-up, etc). What I'm trying to say is that it seems like we were doing everything at the same time, and yet nothing in what appears to be the traditional sequence. What should I think of this?
I know exactly what you mean, and share your pain! When I first started to study lean, I watched Toyota advisors help a supplier to improve their lines. They seemed so certain of what to do next at every improvement that I was convinced they were following a "roadmap." I kept badgering them about it, but they always denied it. They claimed they were simply focusing on problems as they appeared. It took me years to actually accept what they were saying. That’s because it took so long for me to understand what I was seeing.
The best way to clarify this is to go to the Gemba. Since we are discussing sequence, let's consider what usually prompts a lean intervention. An area manager working with the plant manager or the sensei will look for problems such as:
- The cell creating the most quality issues.
- The equipment with the most frequent stoppages interrupting the flow.
- The most inflexible process generating the largest inventory.
- The worst organized activity, carrying a lot of waste.
If these issues are immediately evident, then the plant manager or the sensei will ask the team leader to improve the area by choosing the one that best improves his or her understanding of their processes. This area manager must work with her people to 1) grasp the situation thoroughly, 2) try various fixes with kaizen, and 3) come to a fundamental understanding of the problem and establish new standards for this process so as to solve the problem completely.
If issues aren't immediately apparent, then the sensei will probably say there is too much inventory in the system, recommend that stocks should be lowered in order to make the problems appear (lowering the water in the river to see the rocks), and then pick a challenge for the area manager to solve.
Problem with Traditional Scheduling
Unfortunately, for this approach to work on the shop floor, the lean team must have already gotten the process into more or less just-in-time conditions. This is essential for lean practice to gain traction - for individuals to see problems as they emerge. When stocks are not physically located at the right place (at the end of every process which produces the parts), or if parts are not pulled through the process, then lowering inventory levels will not reveal problems! Most scheduling systems are in fact designed to hide problems. They are no more than built-in workaround processes. Traditional scheduling systems automatically re-route work to machine B if machine A is down for repairs (hence creating an inventory of machine B parts) in order to keep the operators of machine A occupied and so on. Lowering the inventory in this environment does create problems, but it doesn't reveal them. The system just stops delivering, and it's almost impossible to figure out why and how.
In most traditional environments, people are so busy just making the parts and shipping the parts that they reject the very idea of an ambitious improvement project. Individual problems are hard enough to deal with; so fixing the entire process is just plain overwhelming.
The biggest challenge in launching lean is establishing the basic lean principle of managing by problem-solving - of developing people through getting them to understand their problems as a means of improving their processes and their work.
Experience shows there are roughly four main steps to help people understand, and practice, this problems-first approach:
- Clean the window
- Fix quality issues
- Get to rough-and-ready pull quickly
- work on lead-time reduction projects
When you step on a new Gemba, you will often find that few processes are visible and that few if any individuals have time for kaizen (these two conditions are in fact interrelated). Because we're all human, the managers and people working there will agree they have many specific problems (mostly due to lack of investment), but will deny forcefully there is anything fundamentally wrong with their processes or the way they manage the plant. The first issue is to find a way to change their attitude - to see their problems and to accept new approaches.
Typically this is where targeted kaizen events are useful. Flow and layout workshops will typically reveal productivity potentials of about 20% to 30%. Five S workshops will reveal the conditions that equipment is in. Although each of these isolated events will deliver spectacular local results, they're unlikely to affect the entire process, or even to be sustained over time - that's because the root cause of the problems will not be resolved, so the rubber band will pull back the process to where it was. These events, however, are very useful in 1) getting people on the Gemba to acknowledge they have problems and 2) getting management to grasp the potential improvement if these workshops were replicated on every process.
The next step will then be a wall-to-wall attack on quality issues. This will be typically done by installing red bins for defective parts and self-measured tracking of bad parts and rework in every process. All the other disruptions the shop floor has to endure shift after shift should be tracked as well. You can do this by creating production analysis boards that accompany the red bins, and are designed to compare the expected production to the actual production on an hourly basis, with a space for writing down the main causes of the gap. This in itself won't fix anything, but it will bring frontline managers and employees one step closer to acknowledging their frequent problems. Typically, a management team review of the production analysis boards and red bins each shift will highlight obvious problems and deliver low hanging fruit results, which are great for boosting the confidence of the Gemba in the lean process.
Keep in mind, however, that we've still not solved any deep issues. What we're now doing is containing problems by brute force and tighter management control. This will only last a limited time, as people naturally exhaust themselves. This is like putting sand bags in the levee without fixing the breach. There's always more water coming, and at some point, people give up and let the system drift to where it was before.
This is why the third immediate step needs to be to install a rough-and-ready pull system, as an architecture to kaizen. The pull system doesn't need to work completely. The easiest place to start is to focus on the 10% of parts that are the highest runners (typically, these will represent 50% of the cumulative volume). These parts should have a level plan for the week, the components can be placed in "shop stocks" at the end of every cell, and they can be scheduled quite easily with a kanban stock replenishment system. Such a pull system will be rudimentary and in no way address every case, but it already will take some of the onus for visualizing every problem away from frontline managers by making it systematic. At this stage, problems will start to appear in the course of the day as the pull system stutters and falters. Typically, there is also a quick hit as installing the bare bones of a pull system invariably makes you realize you're holding about 20% more parts in stock than you need.
Once the pull system is firmly in place, it can be continuously refined, but typically, problems are now revealed where they happen: the process is now in basic just-in-time conditions and real lean work can start.
At this stage, we're in a position to do things right and get managers to solve one issue at a time. These problems can be framed in terms of lead-time issues:
- Improve the quality and lead-time of delivery to customers (this is the stock-to- customer or end of production-to-customer loop).
- Improve the quality and lead time of the production loop (this is from order of this specific product to delivery to customers).
- Improve the quality and lead-time of the supply loop.
These projects are true lean projects. Starting with a detailed materials and information flow analysis, the manager can then learn to visualize the process on the shop floor in more detail, grasp the situation, and engage people who work in the process in seeing, discussing, and solving issues by trying things out. The manager can then learn to improve her policies so that the problems are solved, inventory can be decreased and lead-time improved. In the process, she will also learn to do this with fewer resources, while delivering better quality and on-time delivery to her customers. She is learning to lean her processes.
Standardization, by the way, occurs all along. The very first flow-and-layout workshops will be the opportunity to experiment with standardized work (set sequences of steps to work at takt time), just as the early 5S will be a platform to create work standards. Certainly, fixing quality issues will lead to a significant effort in updating or creating works standards and so on. Standardization happens all along this learning process - indeed, work standards and standardized work sheets are essential of the learning process itself.
I believe that your experience with Toyota veterans differs from your TPS studies because the literature typically describes a plan of implementation of lean solutions and techniques. My experience in working with senseis is that they do have broad roadmaps in their minds, but these are more about the learning path itself rather than specific technique implementation.
They will first make you acknowledge you’ve got a problem. Then they'll get you to visualize "normal conditions" across your Gemba. They'll get you to work on the most obvious (to you) problems that appear then. Then they will push you to get to rough just-in-time conditions because this gives them a tool to work with for coaching you in fundamental improvement. Once this is in place, they'll pick projects for you to learn more about your own processes and how to lead your people to improve.
My experience is that the projects they pick for you are rarely to your liking (why the hell do they want me to worry about this, I have far more pressing issues…). This is about what they think you should learn as opposed to what you'd like to learn. Broadly speaking, the journey is about getting people's attention first, then "cleaning the window" or "clearing the clouds" by focusing on quality, then getting into rough just-in-time conditions, at which step the real lean work can start in the form of challenging kaizen projects.