If you don’t include some measure of employee satisfaction, retention, wellness, etc, then you will optimize the system at their expense. As awful as this sounds, it’s no different than operating a piece of machinery above its capacity, or without adequate maintenance, because doing so improves some short term measure. If you can absorb the cost of doing so, then you might happily burn through lots of machines and that might be acceptable. But people aren’t machines, and you will either find yourself having retention issues, hiring issues, or like Amazon, literally running out of people willing to work for you even as they literally kill their employees.
The most important thing I take from my reading of Deming, is that there comes a point when you have to put the charts aside and deal with people. If you push people over and over and over, eventually they’re going to snap. Even if they don’t, can you live with yourself at the end of the day?
You also have to consider more factors than just P&L. If you push sales people to the point they start breaking the law, or you push machinery and operators to the point of a massive oil spill, you could jeopardize the whole business, or worse. You need to know where the limits are, and that only comes from understanding the people, and the process, beyond just the charts.
I call it the Soylent Green Principle: companies and other organizations often try to pretend otherwise, but they're actually all made of people. I have to remind myself of this regularly.
I was literally referred to as "a resource" by a big company client at a previous job. Like, "once your resource [i.e., me] is on site, let's...". This was years and years ago and it still stands out in my memory.
I hear this a lot and it really doesn't bother me. It's just an abstraction over the concepts of employee, contractor, seniority, team, etc. One resource is roughly equal to one person, but can also be half time each from two people, so "person" doesn't work.
It's a little odd to use the abstraction when you know exactly the person you're dealing with, but like any jargon, people who live it all day end up using when vernacular is called for (e.g. the electrical engineer who tells their spouse that the dishwasher's heating element has become too resistant).
My province in Canada has treated healthcare workers like absolute shit and now the leaders and the people who voted for them are genuinely shocked that those healthcare workers are leaving the industry or finding jobs elsewhere.
It's absolutely insane because you need these people to train the next generation of healthcare workers which isn't going to happen now because they're leaving.
The UK has for a long time "solved" this by importing healthcare workers from abroad, contributing to exacerbating the same problem and driving up the average cost per staying healthcare worker for a whole range of third world countries... Can't blame the people coming here who see significantly better opportunities in the UK, but definitely can blame the politicians not doing enough to ensure enough are educated here and/or ensure they are treated well enough to stay in the jobs longer.
ON, SK, and other Conservative Party-led provinces are seeing the same thing, or moves in that direction.
Make the system shitty so that there is a justification to privatize. AB is just the most balls-out about it since Danielle Smith is basically a female Bill O'Reilly who got elected to Premier (aka Governor for the USA-ians).
Boomers dying earlier than their parents, and stressing the system in large numbers, are also stretching these systems. Even if a NDP-led-double-down-on-healthcare approach there would be serious issues due to demographics.
That makes healthcare cheap, though. We have the same in France, doctors have quite a low pay (of course in the top of the locals, but not as much as even a startup, and startups don’t advance humanity), nurses have a really low pay, some can’t afford living nearby. But on the upside, I barely pay 26€ to see a doctor.
It’s the heart of convincing people to work hard for very little. If I were the most gruesome capitalist ever, I’d be a socialist. The TikTok dances are nice, though.
Everything is advancing humanity, then. Selling weapons might be seen as advancing humanity.
Startups aren’t particularly human in average, neither in their goals nor internally.
It’s just that, at the beginning of a new industry, all the low hanging fruits like “Register everyone in a big directory with photos”, “Make a video app where you can learn things” seem like advancing humanity, because it’s a leap forward. But they’re no better than in a mature industry, doing proper business.
A doctor, a nurse talking with a dying person is advancing humanity. You, noticing a depressed man in the office; a kid talking to the bullied kid in class, that’s advancing humanity.
No, Jira isn’t advancing humanity, no more than Carrefour negotiating rebates on vegetables against farmers.
Just that there's a spectrum, and because most things are very close to zero (and because there's a lot of small positive things) it doesn't mean that everything is above zero :)
And ... selling weapons might be positive might be negative. Selling weapons to defend against Russia might be positive right now. Selling weapons to dumbass US civilians who then can't keep it safe seems negative.
How is talking to anyone dying actually advancing humanity? Hitting the brakes on the cult of meaningless clinging to life in hospitals would be some advance. Allowing people to say goodbye while they still can...
Jira definitely is from hell, but when it was not this cruel cloud-crazy incumbent it was doing some things better than the alternatives back then.
I'm not saying startups are good because they are small/new/risk-taking, but I'm trying to say that incumbents are usually bad because they have more monopolistic tendencies (than the startups that want of course also usually want to eat the market).
Indeed. Capitalist: this pie is the whole world, and I own it - you want some, pay up front. Socialist: I want a piece of that pie, damn I helped you cooking it, only for scraps - you can't have it for yourself. Me: a whole universe subsist within and around that pie that nobody care about, let's enjoy that.
He means they drive decisions based on some arbitrary, often flawed, short sighted, partial picture metrics (revenue, number of clicks, all kpis) forgetting that there are real people with ambitions, talents and feelings that are not accounted for behind it.
One of the worst cognitive biases which humans suffer from is the belief that observations about individual people can be generalized. How much faulty thinking, including racism and sexism, stems from this simple error? An extension of this common mistake is thinking of people as interchangeable, or fungible, when planning work - pretending who does the work doesn’t matter.
A pattern I've repeatedly seen is companies will plan out the next few quarters of work and then assign some number of people, one person for one quarter to a project, two people for three quarters to another, etc. This process enables executives to make decisions about costs, priorities and project funding. This works for it's purpose, however, the roadmap creation and review process perpetuates the fiction that people are interchangeable, that people can be generically abstracted over.
> An extension of this common mistake is thinking of people as interchangeable, or fungible, when planning work - pretending who does the work doesn’t matter.
I agree with your whole post, but there is a nuance here that's important. This fallacy becomes true in a project large enough to average out individual variance. If you need 5000 software engineers, you probably don't care about their individual traits (you might care about the selection process and aggregate traits, but that's another level up).
So at a certain level of executive, this fallacy serves them well. And it probably serves their reports (who might manage 750 people) decently. But their reports, managing 150 people, hear the language and see the behaviors of their bosses and emulate this belief in fungibility, and it really breaks down at this scale. Even worse for the next levels down, who might manage 30 and 5 people respectively. Yet everyone who sits in at an important meeting of skip-levels and execs sees this model used for resourcing.
Organizations are built with people but often made of pure assets: machinery, software, capital, databases and datalakes, business rules make up the soul of the business. And eventually with GPT, you could feed tons of business data to the AI and the business itself will begin giving instructions for what markets to expand to, who to hire and fire, what changes need to be made, how to set prices.
Organizations are not just people. Those days are a distant memory. The people will be the most replaceable components, due to their ephemeral nature.
Customer lock in, hubris about the superiority of their product or service, IP or reputational dominance in the industry... lots of reasons leadership can treat the people as expendable cogs who should either shape up in gratitude for their salary, or ship out.
How many people do you interact directly with when ordering something from Amazon? (Not to pick on them specifically, they're just an example, hardly unique in this way.)
Some companies must treat their employees well in order to maximize profits, and others are better off doing the opposite. It’s more a question of the overall market than ‘core values’ or other such BS.
Even within companies you can see both approaches used. Disney animators are treated well even as they’re worked to the bone because they can’t get away with actually mistreating them. Other employees aren’t nearly as lucky.
> Even within companies you can see both approaches used.
Every company I’ve worked for has been a great place to work for some categories of people. If you weren’t in one of those preferred categories, your treatment could be anything from okay to grueling overwork depending on how easily they thought they could replace you.
As engineers we’re often lucky to be on the most generous end of this spectrum, especially in recent years. However, I made the mistake of joining a company where the CEO believed having engineers was an “unfortunate necessity” to run the business. It went about as well as you’d expect. Unreasonable expectations, demands, working hours, deadlines, and a constant groveling at our salaries.
IT admins are an example of chronically under-appreciated employees. They keep things running and supported in an office, but I’ve seen them treated as disposable in too many offices.
If a company is a terrible place to work for _some_ class of employees, it will eventually get around to being a terrible place for your class of employee. It's just a matter of time.
Monopolies compete with workers outside of their industry/location.
People working on oil rigs take serious risks, but they also make a lot of money relative to their skill set because otherwise they couldn’t get anyone to do the job.
Training generative models for motion is much harder than independent static images. The training sets are smaller, and the output needs to be consistent over time.
These models might be useful for generating character concept art, but that’s a tiny fraction of the overall workflow.
Yea, they used actors to maintain consistency while manually animating quite an bit, and it still looks janky.
It’s an interesting process but more trying to include X in a workflow than a significant advantage. Imagine trying to do a flight scene where characters are tossing fireballs at each other with that workflow.
That said, it’s very possible for this to be used on a specific movie. The pick a plot that works with the limitations and let the jank be part of the art stile…. I doubt you could watch it for an hour without getting headache, but used sparingly it could work.
I admit I don't know much about that field. But I don't see why video generation methods not follow similar rate of improvement as image generation. There's almost infinite amount of training data (e.g. YouTube), and the hardware keeps improving. I can easily imagine - in a couple of years - typing "generate a 30 seconds video in style X where characters are tossing fireballs at each other (optional description of the characters and the scene)" and getting professional looking results. And in another couple of years generating entire movies.
It’s a question of internal consistency. The video you linked skipped that by starting with actual actors, but “generate a 30 second video” means no shortcuts.
There’s a common issue in AI generated still art for dynamic shots like someone dunking a basketball to have extra arms. This doesn’t show up in pictures of people standing still because the training set is completely dominated by people standing with their arms to the sides. But in dynamic shots either seems fine and it doesn’t understand the underlying limitations the way a human does.
I don’t think it quite needs AGI to “make a 30 second clip of a fireball fight” but it’s much much harder than a still image of a “wizard holding a fireball.”
> If you don’t include some measure of employee satisfaction, retention, wellness, etc, then you will optimize the system at their expense.
I would go further and say: if your process cannot tolerate the things you should expect when working with humans (e.g. someone needing to go to the toilet, someone staying home sick), it is just a badly designed process.
As you said the cost of overreaching will show somewhere else. Maybe not immediately, but it will. And if you are unlucky it will show up with a literal bang and blood on your hands.
When I read this, I ask myself, why don't people usually work as hard for themselves as they do for these exploitative systems? Large companies may be like giant insects sucking out your life energy. However, I've noticed that people often manage to get up in the morning for these 'insects' but struggle to do so when it's about themselves.
Dealing with the uncertainty of starting your own business is even worse.
Few actually make it and most who try and fail are pretty miserable later on.
The daily grind is also miserable, but it's a predictable misery.
I do side gigs occasionally and if I were to live off this, I would fail - first and foremost because until recently I was too afraid to charge people amounts that make it worth it for me.
The default human brain certainly thinks as you do, otherwise, people would exhibit the same or greater work ethic when it comes to their own stuff. However, I genuinely wonder if this is true.
The saying "a bird in the hand is worth two in the bush" seems reasonable at first glance, but it could also be something someone tells you to prevent you from seeking a better place beyond his influence.
I don't think anyone would argue that having the resources of large corporation gives the average worker capacity to be significantly more economically productive than they would be working for themselves - meaning there's fairly few people who could earn enough to make a decent living self-employed. Even if that weren't the case, we're social creatures and a big part of our motivation stems from our wanting to be seen as contributing towards shared goals and working as part of a team: when that motivation isn't there on a regular basis a good many of us aren't likely to work nearly as industriously as we would otherwise. So yes, it's not hard to believe a company that generally treats its workers poorly is still able to make them more economically productive than they would be solo - what's more dismaying is when they seem able to sustain higher levels of productivity from their workers than other businesses that do in fact treat their workers well, enabling them to offer more in remuneration and various other perks. But not necessarily hard to understand why it happens - the former may simply have chanced upon a better business model or got a lucky early foothold in a market and continues to do well based on that. How sustainable that is in the longer term is open to question.
> I don't think anyone would argue that having the resources of large corporation gives the average worker capacity to be significantly more economically productive than they would be working for themselves - meaning there's fairly few people who could earn enough to make a decent living self-employed
I'd say it entirely depends on industry, if it allows you to buy a bunch of expensive machinery that makes work of worker more efficient - solo craftsman can't afford and don't have volume to pay for say fancy CNC router - sure it can.
But on the other end, like programming, generally the bigger company is the more levels of indirection and waste happens, and small team can vastly outpace in per-developer productivity a big team.
Now when that team needs to buy 10 million worth of hardware to train their AI stuff that flips back to "corporate with bigger resources enhance productivity
Do you have a source for the claim that companies that treat workers badly are able to extract more productivity from them? Or is it just that companies that are enjoyable places to work don't need to pay as much to retain workers? Or you mean that "worker oriented" companies don't prioritise productivity at worker's expense? Maybe I misunderstood your point there.
Edit: you also have to take into account shareholders sucking the value out of companies, generally companies pay as little as they can get away with rather than the opposite
I didn't make such a claim in general, only that it's not so hard to see how there can be such cases. Generally I'm skeptical that it makes long term business sense to either underpay or disrespect workers.
> why don't people usually work as hard for themselves as they do for these exploitative systems?
Choice.
They don't have a choice about getting up and going through the M-F slog. They have to so that they can afford to live.
For personal projects, people have choice and can weigh in all of the other things their lives need. Including recovery time from the M-F slog. Their contributions to personal projects are not coerced.
I like your analogy, though some companies don't treat employees like machinery which needs to be maintained but instead as consumables that just need replacing when they wear out.
That's when there's no incentive to measure employee happiness, much as there's no incentive to monitor the mood of a ream of paper.
> You also have to consider more factors than just P&L. If you push sales people to the point they start breaking the law, or you push machinery and operators to the point of a massive oil spill, you could jeopardize the whole business, or worse. You need to know where the limits are, and that only comes from understanding the people, and the process, beyond just the charts.
So what? Catastrophic, disastrous failure has long since been priced in. Remember the Exxon Valdez, one of the worst environmental disasters in history? Almost completely forgotten and to my knowledge no person was held accountable other than the captain who had to pay a 50.000$ fine.
As long as C-level execs can drive a company to the ground with completely no personal repercussions, why should they not do anything to maximize their ROI (=their bonus payments)?
It prompted a lot of discussions in my workplace cafeteria and I recall it was discussed at length on Blind. The consensus I saw at the time was that Amazon has a really rough culture and something needed to change.
I have not worked at Amazon, so I cannot speak directly to any of this, though.
Workplace conditions at Amazon (both tech and non-tech) are quite awful. There are countless reports of workplace bullying, harassment, and retaliation[1][2][3][4]. Regarding PIP/Focus, managers will make things up and falsely report status to HR. It's an incredibly toxic experience. Amazon also uses its tremendous power to do things like execute unlawful surveillance on current and former employees, and falsely accuse them of crimes[5]. Can you imagine trying to raise kids while Amazon is trying to take everything you own, without legal cause? Unreal this is allowed to happen.
Regarding "death," warehouse workers are seriously injured twice as much as workers in other warehouses[6]. Amazon received numerous citations from OSHA last year for reasons including underreporting and mis-classifying incidents involving serious injuries and illness, as well as withholding information to OSHA. It's gotten bad enough even shareholders are pushing for change[7].
> It prompted a lot of discussions in my workplace cafeteria and I recall it was discussed at length on Blind. The consensus I saw at the time was that Amazon has a really rough culture and something needed to change.
There have been multiple reports in the Seattle Times about how brutal of a place to work for. It's been a while but I recall a headline about how often an employee broke down and cried at their desk.
I could not find any evidence of Amazon literally killing an employee in that article. What I did find was this:
> claims that Amazon failed to warn employees of dangerous weather or provide safe shelter before a tornado slammed the Edwardsville facility
Such behaviour certainly warrants a lawsuit, and I hope Amazon gets into deep trouble for that, but what killed that employee was a tornado.
Now the meaning of "kill" would probably include using force to have an employee being hit by that tornado, but the article doesn't seem to mention anything like that either. The details are left out, so what most likely happened is that the employee was threatened with being fired (explicitly, implied, or at a different time) if they don't continue to work, so they did, and then got hit by a tornado.
So to clear this up: Are we actually talking about a case of manslaughter happening at Amazon, or is this a weird way of pointing out their abysmal workplace safety?
Amazon built the warehouse. Amazon hired the people and ordered them to work in that warehouse. Amazon failed to warn employees of dangerous weather or provide safe shelter before a tornado slammed the Edwardsville facility on 10 December. McEwen and five others were killed in the disaster.
If a mining company doesn't warn their miners of an earthquake, and it kills their workers, it's absolutely the mining company's fault. So how can this be anyone else's fault, but Amazon's?
> So to clear this up: Are we actually talking about a case of manslaughter happening at Amazon, or is this a weird way of pointing out their abysmal workplace safety?
Is there really a difference? The definition of manslaughter depends on the juridistiction, but I think this is clearly a case gross negligence manslaughter, by Amazon.
I am not a lawyer, I don't consider this particular question (whether it's legal to kill your workers or not) relevant to the question of whether they're responsible for their death (they are), and I don't consider lawmakers as the ultimate authority of ethics.
By the quote you posted, I was trying to imply that I'm speaking outside of any particular framework of law, but rather of ethics. Perhaps I shouldn't have chosen the term used in law, but I didn't know what other term to choose.
Law comes from ethics, anyway. It's an imperfect reflection of the ideal of justice, and in this case, it is my personal opinion that what Amazon did should be considered gross negligence manslaughter.
To be frank, it's difficult to believe 'gross negligence manslaughter' is the default term you normally use in lieu of other choices. Or that you were unable to think of other terms.
> So how can this be anyone else's fault, but Amazon's?
Why are we even discussing this? I never asked whose fault it is, and actually asserted from the beginning that I think it is Amazon's fault. Such working conditions should not even be allowed by law.
What I did ask for is evidence that "Amazon literally killed their employee", which nobody has provided yet, and probably never will, because it is becoming very clear that it never happened, and that this is just a case of mis-using the word "literally" for dramatic effect.
Amazon is not a person, and cannot kill someone by raising its hand and stabbing them in the spleen. So, what evidence exactly are you asking for?
Are you implying that this word game ("Ha! Amazon didn't literally kill him by a physical attack!") somehow changes anything in terms of blame?
Or are you just arguing semantics for the fuck of it?
> this is just a case of mis-using the word "literally" for dramatic effect
No, this is you not understanding that words have different meanings in different contexts. Amazon has literally killed them, as in, performed a manslaughter. The fact that it is not a person only changes the mechanism - instead of stabbing them, they performed gross negligence. Why you consider that a "literal manslaughter" must have an element of physical assault is beyond me, but in the end, it's just your own definition of the word.
>Amazon is not a person, and cannot kill someone by raising its hand and stabbing them in the spleen. So, what evidence exactly are you asking for?
From a legal point of view, yes Amazon is a person, and that’s the ground on which it can be sued, unlike "Nature" that generated the tornado — though lately things started to move a bit on that side all around the world.[2,3,4,6,7,8,9]
But even that is missing the forest for the tree I think. It seems to me that the point raised by moring is more like can these words be honestly stated as accurate as when The Mail & Guardian titles "The Nigerian government is killing its citizens — again[1]". Although it can be argued that this geopolitical entity is a pure human fiction, there are actual humans that are willingly planning to kill other people.
As awful as the resulting situation might be, I doubt that within Amazon case they were plans and orders to deliberately kill annoying employees. It doesn’t mean Amazon behavior is acceptable, nor is the fact that some government have let Amazon build this situation.
> It seems to me that the point raised by moring is more like (...)
Yes, that was what I meant. It was admittedly fueled by the confusion when I started from "did I miss some major news", as well as a lack of knowledge of what "literally" means, as indicated by the other responses.
> I doubt that within Amazon case they were plans and orders to deliberately kill annoying employees.
I don't see anyone here say Amazon murdered (deliberately killed) people, but arguing they killed people. Kill can mean murder, which implies intent/doing so deliberately, but can also mean manslaughter, which does not require it to be deliberate.
>I never asked whose fault it is, and actually asserted from the beginning that I think it is Amazon's fault
followed by
>What I did ask for is evidence that "Amazon literally killed their employee"
employees died, you agreed it was Amazon's fault, if it is Amazon's fault that an employee died then Amazon literally killed that employee. Notice it was never said Amazon is legally culpable for the deaths of employees which has a different meaning in English than 'literally killed'.
>this is just a case of mis-using the word "literally" for dramatic effect.
I must admit I find the grammatical high-horseplay evinced by the last bit rather weird, for the reasons I have outlined.
A trolley is hurtling down a track towards your employees, who are working on the tracks at your direction. You have time to warn them but do nothing. An employee gets hit and dies. Did you "kill" them?
We're descending into semantics. But I think many justice systems would classify the above as some type of homicide.
Nowadays "literally" is often used as "figuratively" in a somewhat ironic way to emphasize a point. In this case, the point being that Amazon's actions directly led to the deaths of workers.
The word "literally" has long meant both "literally" and "figuratively" to the point that the latter meaning is in some dictionaries at this point, and is reasonably used in this context to imply indirect responsibility. If the person above literally meant literally in its original meaning, then I agree it was misused for dramatic effect. if the person above meant to use it in the sense of figuratively, then it was merely ambiguously written.
There's the business model of "hire people without experience, underpay and mistreat them for 1-2 years while they can't move (because bad CV), replace them with new ones".
Don’t forget that it’s estimated a significant percentage of middle and upper managers are psychopaths, so quite literally YES they can live with themselves. I fear you’re giving an oral presentation at a deaf person’s conference.
Operationally lax companies aren't a panacea either. Constantly answering 2AM pages because you can't reliably ship working software wears people down. Am I understanding what is meant by operational rigor correctly? Life is easier when you eat your vegetables than eating cake everyday. It's not exciting to practice rigor but it's better than the alternative. You can have processes and reasons for doing things a particular way, and still be nice to your colleagues -- these are not mutually exclusive concepts. I doubt the real reason these places mentioned aren't great places to work is primarily due to operations.
you should read Cameron & Quinn. It's about work culture. IF you have a very rule based culture, it could be completely wrong if you want to create and be adaptable. You will simply not have good relationships then and people will be miserable. That kind of culture is more suitable to work that is all about ensuring rules are followed, like a controlling organ. For software dev, it's poison.
No, "somewhere in the middle" is the wrong lesson to take. They're on different axes. You can have an operationally rigorous, pleasant place to work. You can have an operationally terrible, awful place to work. You can have somewhere that's mediocre at both.
The false belief in a dichotomy is exactly the problem.
Not really. It's the point that one might take from the title alone, but the article doesn't substantiate it.
> And from anecdotal evidence I also know that Western companies with sufficient operational rigour tend to be rather … unpleasant places to work. “Is that not the price of excellence?” a friend said, after I told him of this conundrum. “What, you expect such operational excellence to come for free?”
> An obvious point in retrospect;
It's not an obvious point at all, and the author doesn't try to back it up. The point the article makes is that:
1. Deming himself didn't hold the two in opposition; and
2. It's unclear from any evidence that they need to be, in practice.
So, I worked for about a decade in the semiconductor manufacturing industry, where we used SPC and mentioned Deming a lot. There were some not nice places to work, and they mostly were the ones that used Deming's ideas the least. The system that Deming was counseling companies to replace, was one where you measure results only at the end of the line, and punish/incentivize based on the end result, assuming that the problem is due to the workers, not the system. It doesn't mean that Deming's ideas are infallible, but that's what he was advocating against.
If the author had an actual point about Deming's ideas, he never got to it in this essay anyway. It's reasonably well written, but I kept waiting for the part where he finishes his introduction and gets on to his point (to justify the title), and it never happened. I guess that's the "more soon" we are promised at the end?
Exactly. I have worked on highly efficient work sites where everything and everybody was working like a well oiled machine with enough brakes so everybody could do their best job and short enough hours it could be done ad infinitum.
And then I have seen places where unsuprising basic human necessities like "a person has to pee" or "someone is sick" endanger the whole operation.
Sorry but either you want your process to be resilient and reliable OR you want a pile of burning garbage held together by duct tape that is constantly on the edge of collapse for the price of some miniscule savings that are gone when one such collapse happens.
This is a misunderstanding of Deming. It's about how to organize highly repetitive tasks, such as assembly line production, to get consistent quality. Some of it is about managing people, but most of it is about managing process variability.
Read up on the Toyota production system [1] and statistical process control.[2]
> most of it is about managing process variability.
Yes, and this specific article acknowledged that and stated that this is a niche reflection "about [how Deming's] work goes far beyond the data tools and statistical methods of Statistical Process Control (SPC) he helped introduce to the practice of business — and in many cases, seem in direct opposition to those ideas." As in, this is the thesis statement of the article. Some notes in the article:
> Editors Note: This is an investigative note, and only nominally part of the Becoming Data Driven in Business series. A fair warning — this essay is not intended to be as useful as prior essays in the series.
The author also notes this article will be incongruent with the prior articles in the Deming series of articles. The "official" list of articles in the series don't even list this article:
-3) You Aren't Learning if You Don't Close the Loops — An article about the Plan-Do-Study-Act cycle, sometimes called the ‘Shewhart cycle’ or ‘Deming cycle’.
-2) Process Improvement is Trickier Than You Think — The organisational and political dynamics you’ll face when you attempt to due process improvement.
-1) The Disaffected PhD Skunksworks: A Story About Process Improvement — What it's actually like to due process improvement when the organisational dynamics work against you.
1) Goodhart’s Law Isn’t As Useful As You Might Think — Goodhart’s Law says “when a measure becomes a target, it ceases to be a good measure.” It turns out that the field of Statistical Process Control have long developed methods for protecting against this effect. We look at those ideas, and then examine a particular instantiation of the approach in the context of the Amazon Weekly Business Review.
2) How to Become Data Driven — What does SPC have to say about becoming data driven? The central idea it uses is deceptively simple: in order to become data driven, SPC practitioners argue, you need to understand variation. We talk about why that is.
3) Operational Excellence is the Pursuit of Knowledge — Knowledge is defined as ‘theories or models that allow you to predict better’. It turns out that the mindset necessary for operational rigour stems from the pursuit of ‘knowledge’. The beginning of ‘knowledge’, in their view, is a deep understanding of variation. This piece covers the practical implications of this mindset.
4) There is No Truth in Business, Only ‘Knowledge’ — We look at some of the more philosophical implications of pursuing ‘knowledge’ over truth, particularly in business. As it turns out, assuming that there are no stable truths in business is a pretty good hack.
(Idiosyncratic numbering intentional, negative numbers are the "pre-series" to the series. This article itself was published 20 days after the "final" article of the series."
> misunderstanding of Deming. It's about how to organize highly repetitive tasks, such as assembly line production, to get consistent quality.
No. Deming was concerned with human factors, too. Because they are intimately tied to quality efforts. And that means all humans as agents of continuous improvement, not just as drones in the system.
Since we're bringing up rating and grading, and I'm totally not slacking off when I'm supposed to be preparing a class -
Teachers distinguish roughly between formative and summative assessment. Summative assessment is the kind of thing that goes on your transcript and measures your performance looking back ("Inspection is after the fact, after the parts are produced.") It does have value, but its value is not directly in improving the process or the student's learning.
Formative assessment on the other hand, although it might use measurements or grades, doesn't feed these into metrics or transcripts, and definitely not into bonuses or penalties. Any numbers produced by a formative process are not outputs, but inputs to a conversation that ideally looks at more than just the numbers. That's what you want to improve the process that produces the parts (or the learning) in the first place.
It sounds to me like some of the examples in the page, such as the student who was sent to remedial training after receiving two test scores below average in a row, are a kind of cargo-cult Deming process: you might have measurements and control charts but you're not actually using them the way they were intended.
This is also, by the way, why I am extremely skeptical to say the least of any attempts to "use technology to make education more data-driven". It's not that the idea couldn't possibly work, but I have yet to see what I'd consider a good implementation.
One of the things I’ve thought about for quite some time is whether there are any companies, particularly in the software industry, that follow Deming’s philosophy on performance evaluation. In his book Out of the Crisis, Deming recommends against the type of performance reviews that are commonplace in corporate America. This (https://deming.org/dr-deming-called-for-the-elimination-of-t...) is a link summarizing his views.
I used to work for one that followed 11 of the 14 points and had none of the seven diseases -- then it got merged with another company, management was replaced, and the joint venture follows... one out of the 14, and has the five most serious diseases.
If you find another with double digit points, let me know!
I think the explanation is quite simple. The bits of Deming that is easy to boil down into a course you can systematise delivery of and charge through the nose for verification are exactly the bits that look most Taylorist. And when you're selling into a business environment where most management believes in Taylorism, that both makes it easy to sell and easy to buy into selectively.
What you end up with is bias confirmation: the existing managers think they're buying a system that lets them do what they were already doing more effectively (because what they were doing didn't work very well), and there's enough group-think in the peer group that nobody notices their entire philosophy is being challenged.
I'm currently working for an operationally shit organisation (and it's not a small place all in all). The team I'm in get nothing written down, it's all verbal, when misunderstandings happen (which is often) days or weeks can be lost, we don't have the tools we need, we don't get any training which we are supposed to pick up magically people are working excessive hours without overtime pay, resignations are happening and mine will follow.
If enough people resign the workload on the boss (who is the proximal cause of this, although clearly the rot goes much higher up) will probably cause him to resign. And then the organisation will have an IT system that nobody knows how to use and it will be catastrophic.
This situation is horrible for us. People are crying out for documentation and process. I'll spend some time reading this article more carefully, try to work out whether sweet spot is.
Documentation starts with you. When you experience the feeling "why wasn't this written down?", that is your signal to document it. Don't make it rigorous, rigorous documentation is mostly waste.
One of the more useful forms of documentation, especially for the cost, is notes coming out of a meeting. Keep a bulleted list of topics discussed in every meeting, when there are disagreements (eg in technical approaches) record who argued each position. It can help to have the notes visible during the meeting to get people to clarify their points. Record what everyone says they will do as an outcome to the meeting. Send the notes out afterwards.
Edit to add: double check when you're writing down a commitment from someone. "Josh, I've got you down to defrag the hard drives in staging, is that right?"
I endorse this idea. You can lead by example and make things better for yourself and all those that work with you.
The downside is that your boss may actively discourage you from documenting. You may get in trouble if you take any time documenting when you could be, say, coding.
I find that this is only a problem when you make it explicit that you're doing one thing instead of the other. So don't say "I'll spend some time documenting this". The same thing goes for refactoring. Don't say "I'm going to spend some time refactoring". These items count as the nitty-gritty of software development. Your boss doesn't need to know the details. You just do it because it's the right way to do things. Same goes for unit testing, if it makes sense for your work.
When you experience the feeling "why wasn't this written down?", that is your signal to document it.
It would be the case, were there (even informal) processes for writing documentation. But since there is no documentation the process is that there isn't any experienced people have to humiliate themselves and ask the local gatekeeper about procedures that should have been written down a long time back.
No documentation is the sign of a terminally messed up workplace. Run as fast as you can.
You do not need a process to write documentation. You do not need permission. Start writing things down as informally as possible. A company wiki is good, or in word documents that you put in the shared drive. When someone complains that the documents are not standardized, you give them the work of standardizing.
You can choose to not write down what the gatekeeper told you, but why would you choose that?
The question is: why didn't the previous person document stuff and hand it to the next person, after all you need proper documentation for institutional continuity?
The answer is as always: a messed-up workplace, people wanting to make themselves indispensable or important, and as the new guy you are not going to change that. (Me? I've actually been looked at funny and then written up for asking the gatekeeper's bestie about processes!) Man, that place of employment was messed up.
Perhaps it is my ignorance about the data-driven aspects of Deming's work, but it does not seem particularly paradoxical to me that one might achieve success in business by monitoring and measuring the heck out of things and then reacting accordingly. The trick, I suspect, is in knowing what "reacting accordingly" truly looks like for a given set of numbers. In particular, it sounds like he tried to be highly mindful of what was actually human controlled vs. what was due to a system. Deming's red bead demonstration [0] is an excellent example of how managers conflate the two; in it, he has a group of employees who are required to rigorously follow a strict procedure that is statistically going to have a particular rate of defects, and then proceeds to judge the employees by the defects that occurred during their execution of the procedure. That's an example of how you can do a great job of capturing the numbers and still fail to react accordingly. If your employees are being operationally rigorous and your defect rate is unacceptably high, no amount of reward or punishment to your employees will improve your defect rate (at least, not without inducing them to corrupt their operational rigor, which is presumed to be a defect in its own right). Rather, it is your operations themselves--the system--that must be adjusted!
It seems to me somewhat analogous to what we are seeing with Agile software development and the methodologies and metrics that have sprung up around it; the fundamental ideas seem sound, but their interpretation and implementation (the system put in place by which developers are expected to "do Agile") in many organizations has not managed to retain enough of the original spirit to convey the corresponding benefits.
> The trick, I suspect, is in knowing what "reacting accordingly" truly looks like for a given set of numbers.
I gather (from personal and other's experience) that most managers who are "data driven" just want to see data that matches what they are already planning to do. If you present data that is contrary, it's never fully acknowledged or engaged with. Similarly, thinking about systems and incentives is harder than just telling people what to do and expecting it to be done because you said so.
I think it comes down to Deming really being a master of his trade. It reminds me of Bruce Lee's quote: “Before I learned the art, a punch was just a punch, and a kick, just a kick. After I learned the art, a punch was no longer a punch, a kick, no longer a kick. Now that I understand the art, a punch is just a punch and a kick is just a kick.”.
this is no paradox, it’s merely that if you, “in the words of Martin Buber, replace the ‘I, Thou’ relationship with the ‘I, It’ relationship, you relegate persons to the status of things”, and that’s alienation and it’s bad.
“I don't know how to communicate this, or even if it is possible. But the question of justice has concerned me greatly of late. And I say to any creature who may be listening, there can be no justice so long as laws are absolute. Even life itself is an exercise in exceptions.”
Your rules and procedures have escape hatches and loopholes because the world is more complicated and person-ful than they are. Or you have rules that will be used as weapons.
Hmm, was it a speech-to-text error that insterted Buber's name in your comment? At any rate, know that it is possible to edit your comment to reflect the correct name and avoid the misunderatanding.
For an explanation see [1], which reproduces the Martin Luther King quote this way:
> “To use the words of Martin Buber, the Jewish philosopher, segregation substitutes and ‘I-it’ relationship for an ‘I-thou’ relationship, and ends up relegating persons to the status of things. So segregation is not only politically, economically and sociologically unsound, but it is morally wrong and sinful.”
This is not a story of Japan, this is a story of Japanese "Deming" (Toyoya's) culture brought into a USA factory. And you can clearly see that upping the quality and operational rigor actually made the employees much happier and definitely not more miserable. https://www.thisamericanlife.org/561/nummi-2015
The more I read of Deming, the more I got the picture that Deming just wanted businesses to be run holistically. But - this is the critical part - the employees need to not be morons. They need to be smart enough to understand how the business works, how to do their jobs well, and how their jobs relate to other people's jobs, and whatever the business is producing or servicing. Everybody from the janitor to the CEO needs to not be a moron. And obviously that's hard, because most people are not super smart, and we get dumber in groups, and businesses are groups of people.
My understanding is that a big part of not having performance reviews etc is essentially that you're trying to get away from this problem.
"Those efforts that focus on improving the attentiveness, carefulness, speed, etc., of individual workers — without changing the systems, processes, and methods — constitute a low-yield strategy with negligible short-term results"
If your buisness can be severely impacted if one guy didnt get enough sleep/is dumb, it's not the guy's fault. Make a system so someone else checks the work, or provide him with better information with which to make a decision, or any number of other potential improvements which dont rely on a specific person
One of the grocery stores I go to recently added a low clearance warning bar to the entrance of their parking garage. Fine. The problem is, they added it so it hangs at the same level as the convex safety mirror that allows a driver heading in to see a driver heading out. So now the majority of the mirror cannot be seen or just reflects the warning bar.
"What moron hung this at the same height", I thought. "I oughtta complain to that moron's boss". Because the manager is the one who's going to get the complaint, they need to be the one who is actively, clearly focused on the customer needs and the total quality of the product.
Japanese manufacturers are somewhat famous for driving workers, including engineers and managers to suicide. Conversely, Costco, clearly a data-driven org, focuses at least partly on employee happiness. Measure what matters.
My understanding of In-N-Out Burger, a family business, is that they pay the employees adequately and they receive decent benefits. Morale and camaraderie is notably high, and the restaurants can afford to employ at least two dozen people who just crank that kitchen in high gear all day and night. That, along with tasty food, is why I always preferred to patronize them above other dismal fast-food offerings.
“I continue to wonder at the kind of person who names their business philosophy ‘profound’.”
This indicates a misreading of what “profound” applies to. IMHO Deming meant his was a system of gaining deep ie profound knowledge about a process. He did not, IMHO, call his system profound or his philosophy profound.
My first job was at the world leader in chainsaws. They were operationally rigorous but also a great place to work. It’s actually very nice when things work as they should.
>(I exclude Eastern or specifically Japanese companies from the statement above because I don’t know enough about them to comment; Deming’s influence is largest in Japan. Perhaps they’re operationally rigorous, wonderful places to work? I don’t know enough to say, but I intend to find out.)
Now that would be interesting - as far as I know, the Japanese government has issues keeping companies from being literally murderously exploitative. There's even a term for death by overwork in Japanese; karoshi. Of course, not having first-have knowledge, I don't know if that's an accurate view of the situation there.
But does it have anything to do with Deming? Korea, for example, is as bad, and maybe worse given their alcohol culture; they just don't get any cutesy terms like 'karoshi'. Or consider this recent followup about a famous American bonsai master who learned in Japan, and learned such brutal lessons from his master that he's driven away all would-be American apprentices: https://www.newyorker.com/magazine/2022/11/21/the-beautiful-... Are we going to blame bonsai on Deming now?
I hope to someday create a world where people are valued for the love they can bring to those around them, and not beaten or psychologically abused (like this story or my upbringing) for failing to bring enough value to those in power. I'm not doing well enough to create safe living conditions for myself, let alone those I care about.
Awareness is an important step. If you haven't found them already, maybe there are groups of people who are (or were) in similar situations that can lend some support. It's always better to not go the hard walks alone.
I'm not an expert, but as far as I understand, it's not that the management of Japanese corporations forces their employees explicitly, quashing their protests, in the style of labor vs capital struggle in Europe in 19th century.
Instead it's more about peer pressure and self-representation. It's more like fashion among teenagers: they try their hardest to look cool, fashionable, and trendy, even though nobody (and not their parents or teachers in particular) enforce it. But they badly want to look cool, and feel bad if they don't, and their peers look down on them.
The same way employees just don't leave before their boss leaves, not because their boss orders them to, but because they'd feel silly and rude if they did so, and their peers might sneer at them.
Of course the management does use this to their benefit, and does cultivate these views, again, without outwardly forcing anyone's hand, or even being hypocrites. The management just behaves the same towards their own bosses.
This is true to some degree but it’s extended out beyond that. Many apartments for young workers have effectively non-existent kitchens, fridges, etc. It’s common to leave work “late”, eat on the way home, then wake up “early” and eat on the way to work. This is cultural but is a cause and affect of the work culture.
>Instead it's more about peer pressure and self-representation. It's more like fashion among teenagers: they try their hardest to look cool, fashionable, and trendy, even though nobody (and not their parents or teachers in particular) enforce it.
I’m not sure I understand you here. How can peer pressure be stated as "nobody enforce it"?
Nobody formally enforces it, there's no written rule in the company. Bosses do not order their reports to show up before the official hours, or leave past them.
1. Lot's of processes, measurements and so on that are:
2. Based on flawed thinking.
That is it! It is a common recipe. Make an OKR about something that doesn't make a good OKR, and you have 3 months of guaranteed misery as people figure out how they get what needs to be done, while keeping the number happy. Another example is implementing some process like "Agile" and then finding the stream of work you want to complete has to be paused for 4 days because "sprints", so you context switch to something else because "shift left".
The flawed thinking, I can sum up as a worldview, something like this.
"Unlike all other businesses in the world, we have solved the people management and measurement problem with our perfect system that organizes and scores your work, and the system is a mantra, to be repeated by all employees, and they must follow it, and we won't listen to anything to the contrary even from people highly experienced in their field"
So, in simple words, when business decided that he knows better then employees how employees should work, how measure employees work, and don't want to listen to any complains, and force this view to employees?
To some degree you just defined absolutely any business on the planet.
Upd: wtf with random downvotes, I`m feeling like I`m on a reddit.
The first thing that comes to mind regarding Deming’s philosophy and how it actually relates to quality of life working with software is the idea that when screw-ups happen, don’t blame the individual, investigate and improve the process.
Companies I’ve worked for have all been somewhere on a blame-no-blame spectrum, all have required analysis of incidents, virtually none have been willing to prioritize process improvement when causes have been identified.
FYI: I’m drawing from the section of The New Economics when Deming talks about the paper mill.
The problem — or one of the problems rather — is that these kinds of models assume the thing you're measuring is a static thing, and a property of, say, an individual, rather than something that's changeable, or that the thing to be measured is a potential, and the property of a system.
So for example, there's a fundamental, basic assumption that if you rank people, you are establishing measurements of those persons, that they are objects with fixed properties. Another possibility, though, is that your rankings are measurements of your own perception, individually or collectively as some group, and to the extent they are not, they are reflective of the effects of the organizational system.
The truth on average is probably somewhere in between, or more precisely, it probably varies from individual to individual and organization to organization. But the paradigm that is being described assumes that the evaluators are omniscient, unbiased entities measuring properties of some fixed, unchanging objects.
There's still more issues about the power dynamic that enables this assumption. The problem is that everyone pretends as if it's the first case, when in fact it might very well be the second. Or rather, the problem is that we always pretend it's the first case, except in situations where we feel compelled to be honest about the situation due to friends being involved, or some leverage that's hanging over the situation, or something similar. The erroneous assumptions are held because they maintain the power dynamic — it's a model of how organizational elements are measured precisely because that model, however flawed, forms the basis for accrual of resources and control to those doing the evaluating.
I guess I mention this because I think the premise of the piece is a little off. It's not just that "operationally rigorous companies are hellish places to work" because they're unfriendly, it's that their models of the world are wrong, and they're often not as operationally rigorous as they could be.
> these kinds of models assume the thing you're measuring is a static thing
No
> a property of, say, an individual
Also no
> there's a fundamental, basic assumption that if you rank people, you are establishing measurements of those persons
Deming says that's bad
> Another possibility, though, is that your rankings are measurements of your own perception
That's built-in
> But the paradigm that is being described assumes that the evaluators are omniscient, unbiased entities measuring properties of some fixed, unchanging objects
Exactly not
> their models of the world are wrong, and they're often not as operationally rigorous as they could be.
Yes, they've misinterpreted what Deming was saying. Seems to be a common affliction.
Yes! But! There is a sweet spot where providing some level of rigor is a way for people to autonomously move in the right direction without anyone having to micro-manage (which I feel is less nice to work), the key is allowing room for autonomy - and also remembering that everyone you work with/for/etc are people with entire lives going on outside of work.
Rigor is good as long as it contains Slack. Slack should be built into the system, it should be part of the rigor. Because rigorous systems aren't brittle, they are resilient. And you only get resilience when you are able to adapt to the unexpected.
If your plan only works if everything goes right, you don't have a plan, you have a wish.
It seems that many firms have an extractive relationship with their employees, and adding a Deming loop to that makes them hellish. The US auto companies had a famously difficult time adopting the Toyota Way. Many cultures have this, but the current in the USA runs especially deep.
I think a focus on operational rigour also makes companies fragile. You need some flexibility in the system, and also some amount of "waste" (overprovisioning) to respond to circumstances (threats of opportunities) outside your control.
Microsoft was once something that a manager at a consulting firm I worked with held up as exemplaries in software. Yes, the release-beta-to-the-market windows bluescreens etc.
Amazon seems to be similarly deified by software management these days, but... why? Amazon the commercial website frankly sucks. The product information is bad (compare something like computer parts with something like NewEgg). The reviews are suspect, the search results gamed.
AWS has so many warts and fake not-down even in the core of EC2 / S3 / etc, and itself a TERRIBLE console and not-great API, things that minimal investment should fix but never have.
The fact is that software management is still management and they don't get software. They only respect management, which actually is "successful in market cap" not necessarily the impossible to measure actual good management.
they were there first, that is why they were successful
To be fair they were there with e-commerce and cloud services so many years ahead the competition it is understandable to some extent why their products have so many warts. Competitors could see things that didn't work well in amazon products and improve where it was obvious
They have a moat around their products, in retail they have the user-base and in AWS they have the developer-base (devs who know how to set up things in AWS). These things take a long time to erode away as competition increases no matter how good the competition product is
The most important thing I take from my reading of Deming, is that there comes a point when you have to put the charts aside and deal with people. If you push people over and over and over, eventually they’re going to snap. Even if they don’t, can you live with yourself at the end of the day?
You also have to consider more factors than just P&L. If you push sales people to the point they start breaking the law, or you push machinery and operators to the point of a massive oil spill, you could jeopardize the whole business, or worse. You need to know where the limits are, and that only comes from understanding the people, and the process, beyond just the charts.
Classic map v territory stuff.