Boeing's Quality Problem Starts with Its Management Controls

On January 5, 2024, an Alaska Air 737 Max 9 flight experienced a nearly catastrophic failure when a “permanently” sealed plug door blew out, depressurizing the plane. Fortunately, they had just taken off and were only at 16,000 feet, so they were able to return to the Portland airport without any loss of life. Had the plane been over water without an airport nearby (that particular plane had been to Hawaii 22 times) or at a higher altitude, this outcome likely would have been far worse. Depressurization warnings on three prior flights (one on January 4) had led Alaska to ban it from overwater flights on January 4.

Right now, there are two conflicting narratives on Boeing. When I worked for him at McGraw-Hill 40 years ago, Joe Dionne taught me a major lesson about business. The key to the right answer is being at the right layer of the onion, and a common mistake is to dive in and go deeper when the right answer is to go to an outer layer. This is being demonstrated now at Boeing with more in-depth studies of went wrong and how to fix it. It is a quality issue, but I am not sure that is the root cause.

The current report from the National Transportation Safety Board is that a repair was required on the plug door, and at the end of the repair, someone forgot to replace the bolts. In one sense, then, this is much ado about little. A human being making an error is not uncommon, after all. What is a big deal is that Boeing did not inspect the repair. According to a whistleblower published in the Seattle Times, a contributing factor was outdated software at Boeing, where the repair system did not talk to the system that scheduled inspections. Technology debt is a real problem, whether CEOs and CFOs accept it or not.

In addition, obviously, Alaska Airlines did not inspect the plug door when accepting the aircraft from Boeing. Since it was “permanently” installed at the factory, this is somewhat understandable, especially since a similar door is on the 737-900 ER, which has been flying for some time. My undergraduate engineer training tells me that anything secured by bolts is not “permanently” installed. At least in theory, bolts can loosen over time. and periodic inspection seems in order. However, there is no easy way to do that, as the door is installed from the inside and then covered over.

Thus, in retrospect, however, it seems as though Boeing should have identified the door as a potential point of catastrophic failure and recommended inspection. Frankly, I think the airlines should have done this on their own. It seems pretty obvious the door does pose that risk. Most importantly, where were the regulators? This is not that complicated.

Moreover, there was another quite similar failure on the 737 Max earlier, which did result in over 300 fatalities. To achieve the promised performance, Boeing had to change the engines for the Max, and the new engines were larger. When a 737 lands, the ground clearance for the engines can be as little as 13 inches. With the new engines, it would have been 11. To resolve this, Boeing moved the engines. Unfortunately, this changed the aerodynamics of the plane, creating the risk that the plane might stall if the front tilted up too much (the angle of attack).

A major selling point of the Max to airlines was that there would be no need to retrain pilots. Any 737 pilot could fly the Max. If pilot training was required, penalties of $1 million per plane were in at least some contracts, making this promise very tangible. Boeing has always been the pilot’s airplane, and if the avionics and pilot disagree on a Boeing plane, the pilot wins. This is the opposite of Airbus. (I am told the choice is a coin toss.)

Instead of retraining the pilots (which, when finally done, was a 45-minute lesson on an iPad), Boeing chose to add a new system to correct this situation automatically. The new system monitored the angle of attack, and if it was too high, it moved the tail rudder to compensate for a period of seconds. If the problem persisted, the software kept increasing the time it moved the rudder. If the sensor was wrong, the plane was doomed unless the pilot spotted the issue and shut off the electricity to the system by shutting off circuit breakers mid-flight..

Boeing chose to make a second sensor an extra-price option, so the system was designed to use only one, thus creating a single point of failure. At minimum, that sensor should have been checked before every takeoff, and the pilots trained on handling a failure. A well-designed system would have constantly checked two sensors and detected a sensor failure. As shipped, the very existence of the system was obscured to avoid pilot retraining (avoiding the million dollar per plane penalty), and the fact that there was a new single point of failure not mentioned.

What happened is quite simple: Boeing management made two decisions that improved short-term profitability at the cost of introducing a new potential risk of catastrophic failure. They hopefully believed these risks were tiny, but in no way could they be considered impossible. Moreover, it appears that customers were not fully aware of the risks. Had they been, they might not have purchased the aircraft. At the very least, they would have implemented different operating procedures.

From what I have read, a full disclosure on the plug door would have been like this: “For the configuration you have ordered, as we have on the 737-900 ER, we have adapted the high-density configuration for you by installing a plug door in place of the exit doors not required on your configuration. We will install these at the factory, and you can configure the interior as if it were not there. However, this does create a point of catastrophic risk.” Would any customer have bought with that disclosure?

The full disclosure for the earlier computer problem would have been: “To help prevent a potential stall, we have installed an automatic system to move the tail rudder in response to signals from the angle of attack sensor. Should that sensor fail for any reason, catastrophic failure of the plane could occur. You can buy an optional second sensor to alert the pilots to this risk, but the software does not recognize the second sensor. We do not believe this feature is significant enough to require pilot training.” Again, would any customer have bought with this disclosure?

The root of these problems can be found in the Boeing proxy statement. As required by law, it includes the compensation discussion and analysis section. It is very good at conveying to readers the importance Boeing attaches to profitability and free cash flow. That is certainly the definition of shareholder value. However, many other things matter to shareholder value than short-term financial results. I talk about these issues in my book, The New Technology State.

I joined Xerox shortly after their recovery from their quality issues in the 1980s. Xerox, like Boeing now, was in deep trouble from their quality issues, but they turned things around with System 10. To do this, they did two things: they completely transformed their culture to be quality-focused. This was a religion, but it was embraced by nearly all. Quality metrics became as important as financial metrics. This deeply affected me and the rest of my career.

They also did something else. They abolished incentive compensation except for corporate vice presidents and above and for sales. One executive told me how they analyzed their biggest mistakes and concluded that the root cause of nearly all of them was middle managers trying to achieve their short-term goals. This does not surprise me. I encountered this repeatedly at multiple companies and discussed it in my other book, The Bleeding Edge.

Most people, I have found, take their incentive plans as free license to do whatever it takes to achieve its goals. This is human nature. Not evil. John Tukey is one of the greatest statisticians of all time. He repeatedly warned against the human-nature temptation to take a precise answer to the wrong question over a vague answer to the right question. Incentive plans are precise answers to, at best approximately, right questions.

Daniel Pink has also written extensively on the subject. See especially his book Drive. Research shows that incentive compensation is even counterproductive for cognitive tasks. The problem for any large organization is that even a middle manager can make a decision that creates at least the potential for major downstream costs, which may not be at all obvious to more senior management. This is especially true in any complex system like an airplane.

Boeing made these challenges greater for itself by spinning out its assembly requirements to a new public company with different equity and interests. Those old enough to remember Frank Lorenzo’s move to outsource baggage handling at Eastern Airlines will spot the similarity. Customers actually wanted their luggage returned promptly and intact. Spirit Aerospace could be similar.

Yes, Boeing has a quality problem. But that problem did not arise in a vacuum. As a CFO, I always had a signature policy with a separate trigger for higher-level approval based on maximum financial exposure. I think introducing a single point of catastrophic failure is a decision that would trigger such a review. I doubt the Boeing board ever debated this issue, but they should have. Governance and control is part of quality, perhaps the most fundamental part. No one likes inspection, but without inspection there is never adequate quality.

Focusing on financial results as a driver of poorer quality is not a new narrative for Boeing . Many trace it to the merger with McDonnell Douglas. I don’t know, but from the public statements and filings, profitability and cash flow were the mantra. I am sure no one intended these consequences. Boeing is a great company with great people, but people respond to incentives and signals.

As I watch Boeing’s public posturing now, I see every effort to absolve the board, the CEO, and top management of responsibility. In their narrative, this is a quality problem with processes and procedures dealing with worker errors. If I were the CEO, this would be what I would do if I wanted to divert attention from myself. However, the root cause is inadequate controls starting with the Board and the CEO. So far, the Board has been silent. This is disappointing. A great company deserves better.