Beyond the Template: The Messy Reality of My First Toyota A3

 

I walked out of my mentor's office feeling dejected. I had just given him a progress update on my A3 project. "You do not fully understand the problem,” he said. “Go back and redo the second step.”

At Toyota, I was required to earn A3 certification by applying the methodology to a real-life problem. It took 8 cycles of revision before the A3 was complete. The picture above shows the final document. What is not shown is the many crumpled versions covered in red ink.

The most important lesson I learned from the experience is this: The power of the A3 is having the courage to stare at the broken process until you understand why it broke.

Many leaders fail at problem-solving because they are afraid to look at the "Current State" (the truth). They jump straight to "Future State" (the solution) to avoid the discomfort of the unknown. This bias for action is paradoxically what prevents effective problem-solving.  

The following are my core takeaways from those 8 revisions.  

Be a "Detective": Investigating the Current Condition

Thoroughly understanding a problem is the crucial prerequisite to solving it. However, our natural tendency is to seek the immediate gratification of a solution. You must discipline your mind to override this "bias to solve."

Think of yourself as a detective at a crime scene. You must gather evidence before you identify a suspect. This patience pays off in three ways:

• Jumpstarts the root cause analysis. Investigating the problem yields critical clues. For example, I discovered that my defect was sometimes present before entering a specific process, and sometimes after. This observation proved there were two distinct points of occurrence, and therefore, multiple root causes.
• Improves the quality of countermeasures. When you rush, you implement "band-aids" (extra inspections, retraining). When you understand the condition, you can implement systemic changes that inoculate the process to prevent the issue.
• Shortens the total cycle time. Paradoxically, slowing down is the fastest way to fix a problem. If you skip the investigation, you pay a "rework tax" later when weak countermeasures fail, and the problem resurfaces. 

The Prerequisite: Scoping the Problem

You cannot solve a problem if you are trying to "boil the ocean." The overall goal of my A3 was broad: “Reduce the number of defects per vehicle.” To make this solvable, I had to drill down using a “Problem Funnel”:

The output was a solvable problem: “Reduce the defect rate of Camry driver’s side upper door frame molding levelness.”

Principles for the "Detective" Phase

Once properly scoped, use these techniques to investigate the problem:

“Go and See” - repeatedly. You must watch the process until you truly understand the work. I measured the condition so many times that I eventually developed the ability to tell if it was within specification just by running my fingers over the surface. Demonstrate a learner’s mindset and ask questions of the people who do the work. Be sure to first explain to them the purpose of your exercise. This demonstrates respect and helps prevent them from altering their work due to being watched (the “Hawthorne Effect”).

Validate your data. Data helps prove or disprove assumptions. A common belief was that the chrome molding component was the source of the defect. However, I measured 30 components and found they were within tolerance. Yet, once installed on vehicles, 5 of the 30 had the defect. This data proved that something related to the process was a contributor.

 Look for “clues”. Observe the process carefully and look for signs as to what might be contributing to the problem. Use the “fishbone” categories (Man, Method, Machine, Measurement, Environment). Pay special attention to why people do what they do. Often, what looks like a mistake is a rational "workaround". I found that body fitters would sacrifice molding levelness to fix a more severe condition ('blackout levelness'). The team members weren't being negligent; they were trapped between competing standards. This insight allowed me to rule out the human factor.

Ultimately, understanding the problem is about trading the illusion of speed for the reality of accuracy. By putting on your detective hat and sitting with the discomfort of the unknown, you build a foundation of facts that makes the rest of the A3 process possible. 

What's Really Going On?: Identifying True Root Causes

One might think that ‘understanding the problem’ is separate from “analyzing root causes.” This is inaccurate. Your investigation has already gathered clues that will directly inform your analysis.

Identifying root causes is the core of the problem-solving process. You must traverse the chain of causality downward until you reach the true cause. If you stop prematurely, your countermeasures will not be effective. 

The following are the key sequential steps for effectively identifying root causes:

1. Identify & Validate Factors. List all potential factors. Use your observations and data to rule them in or out. Challenge yourself to “prove” your conclusion – can you provide supporting evidence? Do not hesitate to “go and see” again to confirm. 

This is an example of how spending time understanding the problem enabled a jumpstart on the root cause analysis. I had already identified that Body Fit was correcting a different issue that subsequently contributed to the model levelness problem. I would never have identified this factor theorizing in a conference room. 

2. Drill Down to the Root Cause(s). You must identify the underlying root cause for each of the factors contributing to the problem. The traditional “Why-Why analysis” is a useful tool. Ask “why” until it is not logically possible to continue to do so. “Go and See” continues to be a critical tool. 

When the dimensional accuracy of the door frames was proven to be a factor, I visited the supplier to observe their process to further drill down to the root causes (of which there were no fewer than 4!). Note that for most problems, and particularly complex ones, there are usually multiple root causes.

Tip: Aim for process-related causes, not people-related ones. If you end with "lack of training," dig deeper. Why is the process so complex it requires excessive training? 

3. Validate Your Root Causes. This is the step I see most often missed. You are on a mission of scientific discovery; the root causes you identified are hypotheses. The task is to now validate through experimentation using the following options: 

• The “On/Off” Approach: Can you turn the problem on and off by introducing/removing the variable? For multiple potential root causes, evaluate them one by one, or (if you are trained) a Design of Experiment (DOE) can be used to evaluate them simultaneously. 
• Run a Trial: Create a temporary condition, like a jig or sorting, to suppress the suspected cause. If the problem disappears, your hypothesis is correct. For example, we used 3D scans to select a sample of 30 door frames that were all within specification and then tracked them to confirm the blackout levelness condition did not occur.
• Implement a Temporary Countermeasure. Another option is to rapidly test a countermeasure that mitigates the root cause. Measure the subsequent impact on the outcome.  

Ultimately, the measure of whether you have identified true root causes is the effectiveness of the permanent countermeasures. If they are not, go back and reexamine your Root Cause Analysis.

A Word of Caution: If you find yourself stuck in analysis paralysis, the only way to break it is to step back into the real world. Go and see the process again. 

Conclusion: The Value is the Struggle

If you take away one thing from this article, let it be this: The A3 problem-solving process is not linear. If you are doing it correctly, you will redo steps and iterate constantly. Those eight revisions I endured were not a sign of failure; they were the work. The "struggle" of the revision process is not an obstacle to the work—it is the very mechanism that ensures your success in solving the problem.

Do not fixate on the final, polished document; the paper is just a receipt for the thinking process. Instead, celebrate the revisions. A draft covered in red ink is not a mistake - it is proof that you have the discipline to stare at the broken process until you truly understand why it broke.

In my next post, we will explore how to embed A3 thinking into the DNA of your organization.

For a deeper dive into the mechanics of A3 problem-solving, I highly recommend Managing to Learn by John Shook. This is the definitive guidebook we used at Toyota.

1