Hacker News Reader

Target based metrics, be it budget targets or output targets, are too fragile to be useful in my opinion. There is likely a perfect number that's totally reasonable, ensures good productivity while leaving enough room to be safe and grow. But I've yet to see something like that implemented. I've also never seen targets make a company a better place to work, and I would argue that if somewhere is not a good place to work it's inherently less productive and efficient. Management can always compensate for that with more hours or more humans, but you're bloating your system.

Like retail stores with 8 employees on, where some are meant to handle stock and some are meant to serve customers, but you've given them all sales budgets for the day. The stock doesn't get handled, and you have 8 people not meeting their budgets. All the employees are unhappy about that, stressed, and working against eachother to try to succeed.

Competitive metrics just makes it so you have a handful of "star" people, who are great at undermining their colleagues in order to meet their own budget. Rather than an effective team of good people. So in the example, by the metrics, you have 1 good employee and 7 bad ones, no one is happy, your processes have fallen apart and your team doesn't work effectively together. The only people who get fired are the employees, yet the store continues to underperform.

> Target based metrics, be it budget targets or output targets, are too fragile to be useful in my opinion.

The reported target in the 737 MAX case was that Boeing had promised Southwest a $1 million penalty per plane if retraining was required.

So both Boeing and Southwest created a perverse incentive to make the plane less safe. One or two airline accidents can close an airline, so I guarantee nobody at Southwest really wanted to make that deal in retrospect.

Although Southwest didn't have any 737 MAX accidents and they paid for the deluxe MCAS instrument package, they were horrified to find out that what was delivered, what it was supposed to do, and what was documented were 3 different things.

Source: commercially-rated pilot, followed MCAS fiasco daily. Search for "simulators" in the following link:

https://www.forbes.com/sites/petercohan/2019/03/28/did-airbu...

The reported target in the 737 MAX case was that Boeing had promised Southwest a $1 million penalty per plane if retraining was required.

In context that's 1-2% of the sale price of a MAX (depending on discounts). It's plausible that retraining drove MCAS, but that seems unlikely to me.

deluxe MCAS instrument package

No such thing exists. The AoA annunciator was supposed to be active on all 737 MAX planes.

The AoA Disagree Alert was "standard", the indicator was always (stupidly!) extra, and the Alert in fact (accidentally) was only active with the indicator add-on.

https://boeing.mediaroom.com/news-releases-statements?item=1...

Both planes that crashed didn't have the optional add-on, so ended up not having the Alert either.

https://arstechnica.com/information-technology/2019/03/boein...

Great example. Another is that you may have one "star" employee who consistently notices the thing that most needs doing despite incentives, like jumping on the stock problem, and when review time comes around, they're the absolute worst employee on paper. So you fire them, and things get worse, and there's no obvious reason in the analytics as to what went wrong.

Seems to me that at some point they just should say “it’s good enough” and contend with current growth rates. The paradigm of never ending growth pretty is guaranteed to push companies to the point where they will produce inferior goods. Same happened to Deutsche Bank. I still remember when they were doing well the CEO got it into his head to increase their margins even more and declared that 25% was the goal now. There were a lot of people saying that a bank simply has to take on too much risk to produce these numbers and, indeed, the bank blew up. Some industries simply can’t be run like consumer good companies because the consequences of failure are too severe. Obviously this doesn’t stop people from trying repeatedly. With current executive salaries it makes total sense to destroy a company’s base in exchange for a few profitable years.

That’s one of the reasons why I am against more nuclear plants and running them like other businesses . They will start out doing solid work but over time management will get greedy, reduce quality until something blows up.

You see the same thing with wage-theft in large chain stores. Corporate HQ always says that they don't support it, but then they setup sales and payroll goals that can most easily be met with wage-theft.

I believe all of these generally stem from US law making it impossible to charge someone for being an idiot. So provably contradictory claims like "Hit the payroll goals" and "Don't commit wage theft" can be demanded by management all day.

However, actual liability is only accrued by those who act on those policies.

It'd be nice to see liability, once determined at a lower level, forced to follow the organization chart back to the root decision. Possibly balanced for something like direct reports.

E.g. you manage 100 people, 20 of them committed fraud, you have a 20/100 share of the crime

It's even worse because, unless someone is stupid enough to send an e-mail saying "hahaha wage-theft, they'll never catch us," bad-actors can hide behind the veil of idiocy. They might have been promoted up from these stores, know that wage-theft is happening, but continue on knowing that they won't face consequences for it later.

Same for Wells Fargo. Set impossible goals and then be surprised that people are taking shortcuts. I work in medical and even there they squeeze people to the point of reducing quality. At some point you prefer going home on time over doing a thorough review of the code during (unpaid) overtime.

https://en.m.wikipedia.org/wiki/Control_fraud

I'm talking about something different from control fraud; particularly in the case that it can arise unintentionally. The fact that it can arise unintentionally makes it much harder to combat because there is built-in plausible deniability.

Example of it occurring unintentionally:

1. A corporation sets goals for store managers based upon performance metrics that include payroll. These goals are currently realizable by a competent manager without forcing workers to work without pay.

2. Some fraction of managers realize rather than actually working hard to meet the goals, they can get their numbers up by forcing workers to work without pay. This skews the metrics a bit.

3. The most incompetent managers that aren't committing wage-theft are demoted or fired, being replaced with managers that are either more-competent or willing to commit wage-theft.

4. The payroll goal is now adjusted down because the previous accounting-term's payroll average is lower than before.

5. Competent managers manage to get jobs elsewhere, so the fraction of managers willing to commit wage-theft increases.

6. GOTO 2

Eventually the metrics are so skewed that the majority of managers must either commit wage-theft or lose their jobs. These managers are the only people that will get in trouble if the wage-theft is discovered.

One person squarely at the center of this scandal is Boeing's Chief Test Pilot. Here's his bio:

> O’Donoghue’s military experience includes 12 years of active duty as a U.S. Marine Corps fighter pilot and test pilot. He flew operational missions in the A-4M, AV-8A and AV-8B Harrier aircraft, and engineering flight tests on the AV-8B and F-14 Tomcat. In 1994, O’Donoghue transferred to the U.S. Air Force Reserve where he flew the C-130, C-141 and C-17. While there he commanded both the 728th Airlift Squadron and the 446 Airlift Wing, stationed at McChord Air Force Base, Wash. In 2005, he retired from the Air Force Reserve at the rank of colonel.

> O’Donoghue holds a bachelor’s degree in mechanical engineering from the U.S. Naval Academy.

(http://www.boeing.com/company/bios/dennis-odonoghue.page)

This doesn't seem to conform to your theory that it's "cheaper workers" and business majors that caused these problems: sure, they may have created even undue pressure. But it was an all-American fighter jock doing a lot of the actual lying about MCAS.

To elaborate on the management, not engineering, failure: nice domestic marquee credentials for the most-visible leader, but... How much credence was his input given among other leaders? How high-quality were his lieutenants and the teams they led? This is the kind of issue that only gets flagged from the deep work of your team.

Ethical failures can be structural, not merely individual. I think the problem you rightly identify is that decision making power and ethical responsibility are delegated separately, allowing a kind of ethical-risk arbitrage. This seems to be a failure of regulation in this particular case, or in general of ethics in management rather than engineering.

Ethical failures can be structural, not merely individual.

At what point does some other word become more appropriate?

Some other word than ethics? When it's not an ethical problem that you're describing.

What you seem to be saying (without actually saying it, so I have to guess) is that, because it's a failure of a group rather than an individual, it isn't ethical any more - it's something else. But if the group structure pushes individuals into unethical actions, isn't the group structure an ethical problem?

Or, for the snarky answer: If companies are corporate persons, then of course they can have ethical issues.

In a high-stakes situation fundamentally depending on innovation to provide fair value with outstanding safety, you can not drop the ball or ethics alone will not help you.

You're going to need overall integrity in addition to ethics, from bottom to top, starting with elements that are known to be lost, or recognized as not being adequate, before you have a chance to make it to the goal again.

I can't see how these accidents point to cheap incompetent labour unless you are saying that the product managers, chief engineers, principal engineers are being outsourced.

Are you saying that a plane that tends to tilt up dangerously under certain conditions that is designed to be pushed down nose first by a software with no limit to how much it'll push down depending on a single sensor would have been an acceptable solution if the workers were different? This is a problem with the design and definition of a large system rather than the implementation.

So, I think in this case, whoever had any knowledge of this at any level and let it go through are absolutely responsible. While outsourcing to cheap labour elsewhere can and often does become a problem,these crashes can't be attributed to that. This was a very specific failure with very specific people in the wrong.

Are you saying that a plane that tends to tilt up dangerously under certain conditions that is designed to be pushed down nose first by a software with no limit to how much it'll push down depending on a single sensor would have been an acceptable solution if the workers were different?

With different engineers and a focus on product versus profit the MAX never would've seen the light of day. With the focus on outsourcing Boeing had a gigantic mess on their hands with the 787 (and 777X) and no resources to focus on getting a 737 replacement to market.

That’s a great point but I actually read the article as possibly implying, rather than contradicting, the notion that this is a failure that is bigger than any one engineer.

I read this as: engineering ethics were broken because of a toxic management culture.

One of the largest responsibilities of management is to set the culture of the business. They obviously do not have the skill set do the design, engineering, fabrication, or production. They are people people; they are paid to shake hands, make deals, and cultivate workplace culture. They are solely at fault for a toxic, because of why their jobs exist.

I read this as: engineering ethics were broken but ascribing no reason for it. Thus the engineers were 100% responsible.

This is Capitalism 101, as long as there is no real consequences of ones actions (like polluting, causing deaths...), any Capitalistic company has to make more and more money each year. As the profit universe is not infinite they soon hit the thin smoke wall that is used as ethical barrier and pass through it.

If we start to put CEOs and majority share owners in prison things will start to shift a bit.