CHAPTER 9
Stay on Target
Simplicity is the ultimate sophistication.
—LEONARDO DA VINCI, 15TH-CENTURY PAINTER AND INVENTOR
In hard problems of complex systems, there may be thousands of potential variables to look at, and hundreds or thousands of potential root causes. Your journey towards the root cause is one fraught with distractions that can lead you on wild detours or to disaster. To solve a hard problem, you’ll need to quickly and consistently focus your efforts. You’ll need to take special care to stay on target, avoid distractions, and ensure that you’re simplifying.
THE PROBLEMS OF EXPANSIVENESS
What great problem-solvers do is find the root cause in a methodical way, using what they learn along this path to eliminate vast swaths of possible root causes and avenues of inquiry without actually having to study them directly. I can’t emphasize enough how important and powerful this is. In most problem-solving efforts I see, the root cause analysis step is an expanding step, where many ideas are generated and then tested. This comes from the good intention of wanting to be exhaustive, thereby not missing anything: We might try to think of all possible root causes and test them.
In the moldy food example from Chapter 1, “Stop Guessing,” the team had become paralyzed by expanding on the problem. By the time we arrived, they had created a list of over 200 potential root causes. They had worked through half of them with no solution in sight. In this effort, they had wasted months and hundreds of thousands of dollars, and created brand new problems on the line by testing out these solutions. When we applied a methodical approach of eliminating top-level subvariables, we were able to very quickly eliminate almost all of these guesses in a few days. By the way, the root cause was not even in their original list of 200 ideas.
This drive towards expansiveness is a natural instinct: We’re socially trained to think of all possible risks, and all the ways that something might have gone wrong. If we don’t have a better approach, this is the best one available. There are even a lot of formal problem-solving methods that actively encourage expanding possibilities when problem-solving. As I’ve said before, expansive thinking is wonderful when you’re creating new things. But when you’re solving a practical problem, it’s completely unhelpful. Great problem-solvers instead reduce and eliminate possibilities in order to stay on target. They are masters of simplifying.
STAYING ON TARGET
Recall from Chapter 5, “Dig Into the Fundamentals,” the grass on our lawn was too long. We were able to determine that the height of the grass is determined by how tall it was the last time it was cut, its growth rate, and how much time has passed.
At this point it is tempting to go nuts expanding this tree, as it can feel satisfying or important to learn about something completely. But our job is to efficiently find the root cause. So instead of expanding each variable, we want to see if we can eliminate some of the “high level” variables right at the start. Each one we get rid of has many subvariables that sit below it and we can then ignore.
Because this is a made-up example, we can do some tricky stuff like pretend we have a growth-rate log for our grass. For argument’s sake, let’s say we’ve remeasured the rate of growth of our grass and it’s the same rate every week, so it’s not responsible for our new problem. We cut our grass last week, and we always cut it every week, but the grass has become too long anyway. So we can eliminate caring about grass growth rate or time completely, and avoid digging at all into the fundamentals that control these variables. This process is shown in Figure 9.1, with red X’s representing out-of-spec variables, and green check marks representing on-spec variables.
At this point we’ve immediately eliminated almost all variables and potential root causes without having to explore them. We don’t have to care about bugs, rain, sunlight, or any of that, and the problem has been simplified. Once we’ve eliminated both rate of growth and the time, then it has to be the case that the grass was too long when we last finished cutting it.
Since the grass was too tall when we last finished cutting, either our lawn mower blade was too high when we cut it, or the grass had deflected by too great an angle when the blade met the grass. Notice again that in the next step, we’re keeping our scope simple, and limited to the highest-level controlling variables (see Figure 9.1).
Figure 9.1: Variables controlling the length of grass at last cut (blank variables not yet measured)
If we stay on target with these two variables, we’ll find either that our lawn mower blade was too high, or that something was causing the grass to deflect (maybe a dull blade, wet grass, and so on).
That’s the beauty of simplifying and staying on target: By eliminating the variables closest to the primary variable, we immediately eliminate many potential root causes, and we can avoid collecting data about or looking deeply into most of the process or system.
Of course for figuring out how to keep your lawn trimmed, this rigor isn’t necessary. But as we move from very easy problems to very hard ones, the total number of variables involved in controlling our primary variable expands. Sometimes people resist this level of rigor: They are used to solving some easy problems quickly by jumping ahead, so they’re tempted to do the same here, thinking it will be faster. It may be faster in “getting you doing something,” but it doesn’t get you to the true root cause sooner. I’d much rather you learn this here than learn it at the school for hard knocks when you hit the wall with a hard problem and everyone gives up on you.
THE POWER OF SIMPLIFYING
Let’s consider a more complex system that may have a root cause buried deep in the variable tree. By measuring and eliminating higher-level variables, you will see that we can ignore almost all variables at lower levels that control the problem. As shown in Figure 9.2 the X-marks represent off-spec variables we’ve measured; check marks represent on-spec variables we’ve measured; and white boxes represent the variables we never had to identify or measure at all, because they control on-spec variables—so we no longer care about them at all.
Figure 9.2: The full variable tree with variables eliminated
Note how identifying on-spec variables high up in the tree cleaves off almost the entire set of possibilities right away: This is how you take impossibly complex systems and simplify them to become manageable and solvable. Many problem-solving methods don’t support this behavior, and I highly recommend using one that does.
Let’s consider instead a potential tree in which we’ve expansively determined all possible variables down to a very fundamental level, without eliminating anything at a higher level. We would have hundreds or thousands of variables to account for.
Imagine trying to measure and wrap your arms around all of those variables or guess at which ones are important: It’s a bit of a ridiculous prospect. Luckily, you don’t have to build these massive, sprawling trees in order to solve your problem. Staying on target means understanding and measuring only what you need to in order to drill to the root cause and find a simple solution.
Structuring your efforts in this way will give you the confidence to have tenacity. A great problem-solver will know that each step they take in understanding and simplifying the problem is getting them closer to understanding the root cause. They won’t panic just because they don’t see it yet, and they will maintain the discipline to not jump ahead when they have a guess. If they get stuck, they can take a step back.
This approach also means problem-solving teams can measure and demonstrate progress before they know the answer. They don’t measure how many guesses they’ve tested; they measure progress by how much they have understood and eliminated or expanded, as they learn what variables are in spec and which are out of spec. This structure allows them to sanely organize their team’s efforts, assign actions with due dates, and set goals they can hold themselves accountable to. Each understood variable, either eliminated or expanded, is a clear sign of major progress.
Let’s look back at the toilet roll shrink-wrapper from the introduction. We knew what we had to look for because we had eliminated all other possibilities. We knew how the machine was designed to control the shape of the wrapping, and eliminated everything but a mysterious force at a certain point on the wrapper that caused the tear. Because we knew that this force must be responsible for the tearing plastic, the mechanic was able to look specifically for it, and then find it. I want you to appreciate that there was no way that guessing, would have come up with “maybe there’s a random loose bolt that’s vibrating its way into the wrapping cavity.” Clear understanding of what was and was not happening enabled the specific focus required to resolve this issue.
When you move down your tree, keep in mind there may be multiple variables contributing to the failure. So if you don’t eliminate all but one at a certain level, don’t panic: You may have to go down multiple branches. Some problems require two off-spec variables to both be back in spec in order to be solved, where little or no progress is made adjusting only one. In one consumer-packaged plant I worked at, the arms that sealed bags closed (using heat and compression) were angled incorrectly in two separate dimensions, and changing only one didn’t have a measurable impact on the problem. However, remember that when people start talking about a laundry list of things that have to be fixed, they are almost certainly guessing. You’ll know it’s not a laundry list when someone is able to demonstrate that these variables are out of spec, and explain physically how they directly contribute to the observed problem.
This iterative behavior of digging in, and then simplifying, is the on-target path to root cause. You’ll know that you’ve found the root cause when you identify, understand, and measure the off-spec variable that is directly in your control to change (like “presence of bolt” or “cross-sectional area of seal flush pipe”). Knowing why it’s out of control positions you to quickly develop an elegant solution. Don’t “solve” your problem at a high level where you throw lots of money at it to make it go away: Stay on target and keep going until you’ve found the simple root cause.
MISTAKES HAPPEN
A few years ago my microwave stopped working. It was completely blank, as if it had blown something internal. So I moved the plug to another power socket to check the power. That did not work. I put it back in the original socket and noticed the ground fault interrupt (GFI) had triggered. I reset it. The microwave did not work but I plugged in my phone to the second socket and there was power. Knowing that the power was now on and the microwave was still dead I concluded that something had fried inside and that event had probably triggered the GFI. So I paid a few hundred bucks for a new microwave, mumbling about how these things don’t last the way they used to. When I plugged the new machine in it did not work either. I suddenly knew I had really messed this one up: What were the chances of two bad microwaves?
I studied engineering at college and no doubt a little extra “expertise” caused me to jump to a false conclusion. In my rush I had never actually checked the microwave in a live socket that I had checked with the phone. The root cause turned out to be a wiring fault in the first socket. I still have the spare microwave in my garage—turns out they last a long long time. Lessons in humility also last a long time.
You will make mistakes and incorrectly evaluate some variables. You need to be attuned to this and be prepared to go back, challenge, and fix your work.
NOW: STAY ON TARGET
Things in life are always distracting you from the task at hand. Pay attention to this in the next day or so. Look for situations when someone is having a conversation about a topic, and people get distracted and end up talking about something completely different. Notice how you’re taking care of one task and then find something else that needs fixing—such as loading the dishwasher and you notice the floor needs to be swept, and then moments later you’re rearranging the closet. Noticing and eliminating distractions will enable you to stay on target when you’re facing down your next complex problem.