Boeing 737 Max Was Plagued with Production Problems, Whistle-Blower Says
nytimes.com
nytimes.com
I wonder how much of this is directly the result of penny pinching by MBA types. I'm gradually coming to believe that non-technical management of technical endeavors is the root of most failure in the engineering domain.
Too often these relationships break down and become adversarial, to the detriment of everyone involved, and of course the project.
In my experience managers lacking domain knowledge make more harm than good but are excellent putting emphasis on adjacent and auxiliary parts they are good at. Anyone put in charge of something they are not competent in would do the same.
That could be why he led with "I wonder..."
1. Meetings to "align"
2. Cracking the whip.
That's it. That's all they have. They have poor insight into creating team chemistry, culture, focused work environments, mentorship, long-term planning, QA processes, sourcing efficiency feedback, etc. - all those things that are crucial to boosting productivity and creating supersonic teams. The people that really know their stuff and can work miracles are people that have been in the trenches for 20 years, regardless of profession. And they hands-down make the best managers.
That's my honest 2 cents, it's a bit of disgruntlement from working in large enterprise environments with engineering and technology.
I would argue that this is the problem unto itself. The idea that a manager can work invisibly "behind the scenes" to create these conditions is the squarest negative stereotype of the MBA - that they are clerical, remote, number-crunchers who don't have an intuitive human grasp of the unique challenges their teams face, because they lack experience where the rubber meets of the road of their organization. I think this Boeing case conforms to this stereotype, as does my personal experience.
But I don't believe they're a scourge, or anything. I quantified it with "a lot" but not "most" ;)
We're a compound of self-organized teams of Product Owners and Engineers.
Our traditional organization takes 10 weeks for feature X, we need at most 2 weeks, averaging at 1 week (10%).
We have the luxury of delivering the exact same product (as a green-field variant of our classical product as a SaaS solution), so it´s quite comparable in terms of scope.
The main difference is there is no hidden agenda of would-be managers, nothing between the customer than a PO who knows what he is doing (i.e. is doing regular A/B tests, customer interviews, involving the dev team as deep as possible to understand customer requirements).
We were awarded a nation-wide award for digital transformation.
No middle-managers, no HR.
You can reach me @ hackernews@disposable-email.ml for additional details if you like.
Summarizing: we got rid of all non-relevant management ballast and are able to deliver features at a pace of around 10x of a traditionally managed line.
Overtime: around 0% with a tendency to dip below 0%.
Addendum: the POs are 70% MBAs, but they are good (i.e. they learnt to deliver as opposed to manage)
My experience is that these lean approaches work well for something low on the severity scale, like an SaaS solution, but can more easily falter with complex safety-critical systems that blend multiple domains (e.g., mechanical, software, etc.) I think sometimes people interpret the process rigor that gets added to critical design to management bloat.
Much of the trouble at Boeing has been explained by various sources as having started with the replacement of managing engineers at a company known to be founded and run by engineers with executive types after the "reverse takeover"[1] by McDonnel. This seems to be as close as one could get to an A/B test and the results are exactly what I would have expected, as technical constraints are relaxed to boost margins and shrink/meet deadlines. Though I admit this is a small sample size and could be confirmation bias on my part.
>Do you seriously think that someone with an engineering degree would for some reason say actually we don't need to meet deadlines, I won't pressure you to finish these planes on time
No, but I would expect a manager with an engineering degree to be more likely to understand the absurdity and danger of building passenger aircraft with missing tools and incorrect instructions, red flag such issues, and push back deadlines to resolve them.
1. theatlantic.com/amp/article/602188/
+That's not strictly true, people can have different personalities at work vs family for example.
You don't need to be an engineer to do any of these things. You really just need to ask questions to people who understand the details and listen to them. A good manager will know the right people to ask these questions to and follow their advice. Good management is not an easy skill however.
Now what if the salesman was human. This human has a feelings and a personality and other things going on in its life. Now what if you have to manage 100 of these salesmen, all with different culture, religion, education and age... and when you put these salesmen together in teams, they all have a different group dynamic. Sometimes you'll get a salesman who will even try to sabotage you. Plus you need to motivate, train, and lead all of them. For some reason technical people think this problem is so easy. (I'm not an MBA by the way)
Also, this is buried further down in the article than it should be.
> Mr. Pierson and his lawyers declined to answer whether he was seeking whistle-blower protection or filing a federal whistle-blower case. He has hired a prominent whistle-blower lawyer, Eric Havian, and could stand to gain monetarily if he pursued such a case.
"Mr. Pierson believes that the production problems may have played a role in the crashes. In both accidents, MCAS was triggered when a vane installed on the plane’s fuselage malfunctioned."
"Workers were completing jobs out of sequence, leading to additional mistakes."
Everything this whistleblower is saying seems consistent with dozens of other stories which have emerged of Boeing's corroded engineering practices.
I've found over the course of my life that money encourages all sorts of people to say all sorts of things.
For example, it seems like they didn't follow their own procedures regarding the system safety analysis as it relates to MCAS:
"Assessed a failure of the system as one level below “catastrophic.” But even that “hazardous” danger level should have precluded activation of the system based on input from a single sensor — and yet that’s how it was designed."[1].
There are also examples of incomplete documentation, etc. that may contribute to a lack of understanding of the overall system. My tendency is to think a culture of schedule pressure may contribute to this.
[1] https://www.seattletimes.com/business/boeing-aerospace/faile...
I mean, the concern would be that MCAS is just the tip of the iceberg, and that similar fuckups exist within the planes due to this management trouble.
There is some speculation in the article that the vane malfunctioned too frequently which may be rooted in production problems.
It makes sense that sensors on an expensive plane shouldn't malfunction so early.
Two planes crashed and the higher-ups knew about the problems and were actively showing it under rug hoping somehow they can fix it without backlash.
There are so many people involved it's easy to distribute blame or blame the entire process. But if the right email can be tracked down, there's going to be someone(s) in that loop that fits the charges for 200+ counts of manslaughter.
You would be shocked to learn how many faults an aircraft can have at a given time and still be allowed to take off. That's what redundancy is for, if everything had to be all-systems-go every time then you couldn't run an aircraft for a week straight. It doesn't matter as long as you have a spare, and it'll get fixed the next time you're in for maintenance.
the issue with the 737 MAX is that there are only two sensors, and so far I don't see that being addressed. One sensor failing should not kill the aircraft, and it should not disable a control regime that is necessary for the aircraft to receive type certification.
Yep. Next time you fly, look through the cockpit door as you board for maintenance tag-outs -- usually red flags with 'INOP' labeling, most likely on the circuit breaker panels above the pilots. The Gimli Glider incident occurred partially because of a circuit breaker tag-out in the 767 fuel quantity indicator system.
Aircraft manufacturers and regulators produce a Minimum Equipment List which stipulates what is required for flight. If you get your hands on one for the plane you're on, you might be surprised to discover what _isn't_ on it and what is permitted to fail. It's heavy, but here's an older American MEL for the 737 family:
http://fsims.faa.gov/wdocs/mmel/b-737%20r54a.pdf
Pages 35-1 and 35-2 offer some interesting reading about the passenger oxygen system, for example.
There was mitigating circumstances though, the procedures were frequently being revised (the 767 was brand new) and previously it had been permissible to take off without a fuel indicator.
None of this takes away from the huge achievement of their landing. Simulator runs show the landing was near impossible. Had Captain Pearson not been an experienced glider pilot its likely they would have ditched in the river, and the historical survivability of water landings are not good.
The MCAS system, which was dependent on those two sensors, was developed to compensate for the degree to which the plane behaved differently. And Boeing underplayed how much the performance of the new plane had changed, and hid the existence of the MCAS system from commercial plane pilots, to the point where it wasn't even in their manuals.
Had Boeing behaved ethically, they would have told the FAA that the new 737 required pilot training. But that would have diminished sales and hurt Boeing's stock price. That's the managerial misbehavior that many of the commenters in this thread are remarking on.
https://www.nytimes.com/2019/09/18/magazine/boeing-737-max-c...
MCAS was not a fix to make training lighter, the aircraft would not be able to get a type certificate without it.
https://leehamnews.com/2019/02/08/bjorns-corner-pitch-stabil...
I doubt that Boeing would have been so gracious to Pierson if they had not already been waist-deep in shit.
"Boeing specifically highlighted in presentation documents the new MCAS compares inputs from two angle of attack (AOA) sensors as opposed to a single sensor in the original, flawed, design. MCAS will only activate if data from both sensors agree the AOA is too high. MCAS will activate only once per high AOA event. Pilots will be able to override it and the system will not activate automatically a second time. In the original design, the system reactivated over and over. The presentation reiterates the crew can deactivate MCAS by turning off the automatic trim system and manually turning the trim wheel."