The myth of the silver bullet

For the last few weeks we've been talking about corporate culture: in particular, about whether you can build a company's culture deliberately, and about how far that culture is implicated when things go well or badly. So it was through a delightful synchronicity that I recently ran across two very different sources which spoke to this topic in rather different ways.

The Patagonia case study

Building a culture ...

The first was a talk given by Carlos Conejo, LSSMBB, under the auspices of ASQ, about "The Patagonia Ethos." Conejo reviewed the outdoor clothing company Patagonia, and explained how they built a corporate culture deliberately and systematically. Back in the old days, when Yvon Chouinard (the founder) first started to make climbing equipment, he told customers they shouldn't expect quick responses during climbing or skiing seasons. Then, as the company grew, they introduced:

• Flexible work arrangements
• Casual dress code
• Flat organization
• No private offices
• Health food in the offices
• On-site daycare
• Transparent communications to employees
• Classes for employees on how to get involved in local, grassroots environmental causes 

100% of the electricity used by the company is from renewable resources. 

98% of the raw materials used by the company are recycled. 

If you have old gear from Patagonia, you can send it in and they will repair it. 

These principles make Patagonia's gear more expensive than that from their competitors, but customers gladly pay the higher prices because they support the company's mission.

Then in 2022, Patagonia transferred all its paying (but nonvoting) stock to the Holdfast Collective, "a nonprofit dedicated to fighting the environmental crisis and defending nature." The voting (but non-paying) stock went to the Patagonia Purpose Trust, "created to protect the company’s values." Chouinard described these transfers by saying, "Earth is now our only shareholder." (Interestingly, Robert Bosch GmbH has a very similar ownership structure.)

All of these steps have contributed to a clear and embedded corporate culture.

... but not a silver bullet

But it's not all roses. Conejo explained that one of the consequences of the company's pervasive informality was that for many years they were very weak when it came to formal planning, budgeting, and performance management. Then when it finally became clear that these activities were needed, they created a home-grown solution that lurched too far in the other direction. For a while, the business planning process took three whole months to plan each year. Partly this is because—in the name of transparency—it engaged all employees at all levels clear across the organization. But many of these employees had no previous experience in (or even exposure to) business planning or the rudiments of project management. So the value of their input was compromised, or else they had to take the time out to learn the subjects they were contributing to. 

Ultimately, Patagonia grew past these problems. They scaled back the planning process while continuing to emphasize openness and the development of their employees. But two overall messages were inescapable. 

First, culture is important but it is not a silver bullet. You need systems too. 

Second, every culture has its own failure mode. There is no "perfect culture"; each one has some strengths and some weaknesses. Which ones predominate is partly a matter of which circumstances the company faces.     

Boeing, again

All of which brings us back to Boeing.

In recent posts* I've suggested that Harry Stonecipher (Boeing President 1997-2001 and 2003-2005) deserves a measure of criticism for deliberately dragging the Boeing culture away from a focus on solid engineering and toward a focus on the economic bottom line. But the second source that I ran across a few days ago was a blog post that provided important insight into that transition. (See "The Myth Of Old Boeing," by Bill Sweetman.) 

What Sweetman makes clear is that Boeing, back in the days before Stonecipher took over, may well have had a solid culture; and the engineers were surely very smart. But their configuration-control system dated from World War Two! By any normal standards, Boeing should have been totally incapable of building airplanes for multiple customers** in the modern day. The only thing that saved them—for a while—is that they had low-ranking employees on the production floor who understood the archaic configuration system backwards and forwards, and who worked around it with heroic effort in order to get the planes built. But these were individual human beings. One by one they got old and retired. And we all know that any system which relies on heroes to get the job done will fail sooner or later.

This was the challenge that Stonecipher faced when he took over the company. Yes, he insisted that Boeing start thinking about the economics of profit and loss. And yes, in the end, it's possible that he went too far. But part of his motivation at the time was to drag Boeing—kicking and screaming—away from a configuration-control system that made factory production pointlessly expensive and mind-numbingly inefficient.

In other words: if it hadn't been Stonecipher, it would have been someone else. The only other alternative would have been for Boeing to collapse under the weight of its own inefficiency.

To repeat the two points above:

1. Culture is important but it is not a silver bullet. You need systems too.
2. Every culture has its own failure mode, and there is no "perfect culture."

For those of us in the Quality business, none of this should be controversial. In a sense, culture is about making sure that all your people are approaching their work in the right way. But Deming taught us years ago that "A bad system will beat a good person every time." That's why you need both. 

__________

* See specifically here and here.    

** A configuration-control system manages how changes or alternatives are introduced into a design. If you sell a single basic product to several customers, each of whom insists on their own unique package of options, you need a sophisticated configuration-control system to keep track of all the variations so that (for example) United gets airplanes tailored for United and not for American. By the early 1990's, Boeing's system for handling these variations was woefully out of date.   

               

How do you fix "culture"?

For the past few weeks, we've been discussing the recent Quality woes of Boeing and Toyota, woes which some observers have attributed to their company cultures. We've considered whether "culture" can strictly be considered a root cause at all. But of course the biggest question is, What can you do about it? At the very end of my post about Toyota, I described one approach that I've seen work, but it was hardly advice! It was more a counsel of desperation than anything practical.

Still, we know that company cultures do change sometimes. It's not common, but it happens. How? What methods actually work? I know a couple, and I'll describe them here. But the first point to make is that it matters critically which direction you are trying to change. Techniques that work in one direction absolutely won't work in the other.

The way down

The gates of hell are open night and day;

Smooth the descent, and easy is the way:

But to return, and view the cheerful skies,

In this the task and mighty labor lies. 

Virgil, Aeneid, Book VI, Lines 126–129 (tr. John Dryden)

If you want to change an organization to focus less on Quality and more on easily quantifiable measures like cost and schedule, the playbook is well understood. This is the mission that Harry Stonecipher set himself when he took charge of Boeing. Mandate strict targets across the organization for schedule and cost-cutting, and then fire people who don't meet the targets. Make it clear that these principles will continue, that they apply at all levels of the organization, and that there is no appeal. Employees who want to keep their jobs will respond by prioritizing their cost and schedule targets, no matter what. Employees who can't do that, for whatever reason, will end up leaving the company (voluntarily or otherwise). So in the long term they won't pose a significant obstacle. 

As I say, this approach is well-understood. I have oversimplified the picture in a hundred ways (and of course I don't deny that every business has to run a profit to stay open). But a rigorous application of this approach or one like it will reliably orient a company around whatever quantifiable measures you choose. And the values of those measures will reliably improve over time. Partly this improvement is the natural result of unwavering focus. Also, there is no measure in the world that cannot be gamed; and so if the incentives are high enough, people will figure out how to game whatever measures you put in place.

The way back up

If on the other hand you want to pull an organization back up to more of a Quality focus, the approach has to be less brutal. Quality is a direct consequence of loving or caring about your work; and people cannot afford to care when they are governed by fear. But this in turn means that the ascent back up to a Quality focus takes time. The key, as in so many other Quality implementations, is Kaizen—introducing little changes in the right direction, so small that they are easy for everyone to absorb ... and then following up with new ones. Again and again.

I saw this done once, gently but very efficiently. A man I'll call Franz (not his real name!) was assigned responsibility for the engineering programs at his company. Franz soon realized that the different concurrent projects all affected certain common modules. In other words, Project A would have to use module M, so they would change it to fit their needs. But then a day later, Project B would change it again, to fit their needs! And so on. In addition, there were a couple of experts in the company who knew areas that no one else knew; when several projects needed their expertise all at once, the work quickly bottlenecked.

The classic solution to this problem is to set up a Change Control Board that reviews all the upcoming work and decides what will be done when. But Franz quickly learned that his organization was more or less allergic to formal structures like CCB's, even when they promised to be helpful.

So Franz started bringing in a box of doughnuts every Friday morning, and asked all the project leads to come join him around 10:00—casually, whenever they could get away—for coffee and doughnuts. "I really don't know what you guys are working on these days, so come have a doughnut and tell me about your work." Pretty soon it was well established that every Friday all the project leads gathered at 10:00 in the conference room next to Franz's office, and talked casually over coffee and doughnuts about what they were doing. And sure enough, from time to time the conversation went like this:

Project Lead 1: Yeah, as soon as we get our heads above water on these run-time problems, the next thing is we have to dive into the Whitzinframmer Module.

Franz: Wow, really? The Whitzinframmer Module? [Turns to Project Lead 2.] Say, didn't you tell me last week that your team is totally rewriting that module right now?

Project Lead 2: That's right. We are.

Franz: Maybe the two of you want to talk about this for a few minutes, so you don't step on each other's toes.

Project Lead 1: Sure, no problem. [To Project Lead 2.] Can I come by your desk right after this, so we can talk for 15 minutes?

And thus, week by week, Franz nudged this group into a fully-operational CCB. He never used the name "CCB." It was always just "Doughnuts with Franz." But step by step the Quality improved while the frustration and bottlenecks diminished.

There's no magic recipe. Every case requires understanding where you are today, where you want to get, and what tools are available to help get you there. Maybe you can do it with doughnuts, or maybe it will take something else.

The key is that people want to do a good job. You can squeeze that desire out of them by subjecting them ruthlessly to the wrong incentives. And if you do, then it may take a while for them to find their way back once your successor sees why it's needed. But step by step it can be done.     

               

Is "culture" a real root cause?

Harry Stonecipher,
Boeing 1997-2001, 2003-2005.

Boeing and Toyota have both been in the news recently, and it's not the kind of publicity anyone wants. For Boeing, the news has been about Quality problems in their aircraft—most dramatically, in the 737 MAX-9 flying as Alaska Airlines flight 1282, which blew out a door plug six minutes after takeoff, at a height of nearly 16,000 feet. (Fortunately, no one was seriously injured. I wrote about this accident here and here.) For Toyota, the news has been about faking test data on multiple models over many years. (I wrote about that revelation here.)

How can these things happen? I trust (or hope!) both companies are conducting detailed investigations, to find the causes and ensure they never happen again. But we in the public are unlikely ever to see the results of those investigations. At the same time, there has been extensive discussion and speculation in the press. And for both scandals, one of the likely suspect causes has been "company culture."* In other words, the suggestion is that Boeing and Toyota, respectively, are just the kinds of places that do these kinds of things.

Phrased so vaguely, the accusation looks like an excuse: "Gosh, I dunno. These things just happen." We can be excused for not taking it seriously. And yet, ... and yet, .... We all know there is such a thing as company culture. You can detect it in little things: Do you address your boss by first name, or by last-name-plus-title? How do you dress for the job? How punctually do you arrive in the morning? And so on. These subtleties affect how you act on the job, and what decisions you make. Is it really just an excuse to think they might have an impact on the actions and decisions that lead to an accident, or to fraud? Why can't we consider culture a root cause (possibly among others)?

What is a root cause?

As you recall from our earlier discussions of problem-solving (especially here and here), a root cause has several distinctive features:

• A root cause really is a cause. That is, there is an objective, logical chain of causation which leads from the root cause to the effect.
• A root cause is deep enough to solve the problem permanently. If you correct it, you prevent any further instance of the problem.
• A root cause does not assign personal blame. The point is to improve the system, not to single out one person. If he could do it wrong, maybe the next guy will too.
• A root cause is actionable. It's a waste of time to name a "cause" that you can't do anything about.

How far does culture meet these criteria?

I think the hard ones are the first and the last. With respect to the second criterion, culture is as deep as you are ever likely to get. If anything, it may be too deep. But if you can find a causal chain that implicates it (the first criterion) then I don't think you need to look any farther. As for the third criterion, culture is surely a system-level phenomenon. Individuals might strive to shift a company's culture in one direction or another—Harry Stonecipher famously (and avowedly**) strove to shift the culture of Boeing from one centered on engineering to one centered on profitability—but everyone participates in a culture. The culture influences decisions that its instigator (if there is one) doesn't know anything about. So it is not a matter for personal blame.

This week, let me unpack the first criterion. Is culture really a cause? Then next week I'll discuss what we can do about it.

Is culture really a cause?

Koichi Ito, President Toyota Industries Corp.

We look for causes by using some method like a 5-Why analysis. And an important way to double-check that your 5-Why analysis has not gone off the rails is to read it backwards from the last cause to the first. Can you say, "E, which caused D, which caused C, which caused B, which caused A" and have it make sense? So to ask whether culture is discoverable as a cause is equivalent to asking, Could we ever find ourselves saying, "Our company culture is like so, and therefore the engineer didn't tell his manager that the design was unsafe"?***

Well, maybe. It's important to remember that 5-Whys uncover more than just direct or proximate causes. (These are sometimes called technical root causes.) In addition, it is always fair to ask, What is it about the way our system is set up that allowed this situation to arise in the first place? This is the managerial root cause, and it can be summarized by the whimsical remark that "Everything that goes wrong is the fault of senior management." (See also the extended treatment of this topic in this post here.) 

For example, "The part is crooked, because the machine slid out of alignment, because management didn't implement a preventive maintenance program." 

Or, "The circuit failed, because the engineer designed in the wrong component, because he copied the design from another product without realizing there is a different use case, and no one caught it because the development process doesn't require a Design Review." 

A full causal analysis (sometimes called a 2x2x5-Why) should include such elements, so it is not crazy to think that a reference to the company's culture might turn up.

Even so, notice that the managerial root causes identified in the examples talk about specific programs (preventive maintenance, design review) that should have been implemented but weren't. The presence or absence of such programs contributes to a company's culture, but the actual causation is tied to the concrete program, and not to the culture as a whole.

What about the fictitious*** example that I invented*** a few paragraphs ago: "Our company culture is like so, and therefore the engineer didn't tell his manager that the design was unsafe"? How would that show up in a 5-Why analysis? I imagine the chain of reasoning might look something like this:

Why did the part break?

It was subjected during normal use to strains that it could not withstand.

Why was it built so that it could not withstand the strains of normal use?

It was built to the design.

Why was it designed so that it could not withstand the strains of normal use?

No objections were raised during the Design Review.

Why were no objections raised during the Design Review?

The engineer feared repercussions for causing a schedule slip.

Why did the engineer fear repercussions?

Our company culture ... blah, blah, blah ....

So yes, in a case like that the culture does finally show up in the line of inquiry. But it's not really a single event or action. It comes in as a general explanation for some other concrete action, namely, an engineer not speaking up about a design flaw for fear of repercussions. And if you wanted to prevent recurrence of the problem, you would start with the concrete instances that showed up in your analysis.

I think this means that culture is less a root cause in its own right than it is a summary of multiple other root causes, including some possible causes that might never actually come to light. And that means that corrective action has to be focused more narrowly than just "change the culture." I'll talk about this more next week.     

__________

* See, for example, this article about Boeing or this article about Toyota.

** "'When people say I changed the culture of Boeing, that was the intent, so that it's run like a business rather than a great engineering firm,' Stonecipher once said." From "Boeing's quality-control process and company culture are being heavily scrutinized after the 737 Max Flight 1282 blowout," Business Insider, Jan 13, 2024. https://www.businessinsider.com/boeing-737-max-quality-control-company-culture-merger-finances-2024-1. 

*** To be very clear, I invented this sentence as a fictitious example. I am making no allegations of any kind against any person or persons, company, or other organization.   

          

And now it's Toyota?

How many of you had "Quality scandals involving major manufacturers" on your industry Bingo cards for Q1 of 2024? Anyone have it twice? It looks like we have another winner. (After Boeing, I mean.)

Toyota Industries Corp. President Koichi Ito, front, bows in apology
at a press conference in Tokyo on Jan. 29, 2024. (Kyodo)

What happened?

A week ago Monday, on January 29, President Koichi Ito of Toyota Industries Corporation admitted in a press conference that "power output data had been manipulated for 10 of its models sold globally." [See also this article here.]

"Toyota Industries Corp. fabricated the data on diesel engines it makes and supplies for the automaker, Toyota said, adding that it will suspend shipment of the affected vehicles, including the Land Cruiser 300 and the Hilux.... The 10 models also include the HiAce, the Fortuner and the Innova. The cars were sold in Japan, Europe and the Middle East, among other markets. The output data rigging dates back to 2017, Toyota Industries said."

Nor is this the only recent scandal involving Toyota:

• Last year, Daihatsu (a Toyota subsidiary) stopped all shipments after admitting that safety tests for most of its models had been rigged for thirty years.
• Back in 2022, Hino Motors (another Toyota subsidiary) admitted to falsifying emissions and fuel performance data for nearly 20 years.
• For that matter, Toyota is still following up on the largest automotive recall in history, first announced in 2015, because some models were built with Takata airbags that risk exploding. (See YouTube video below.)

Is there any good news?

Toyota Chairman Akio Toyoda bows in shame during a news
conference addressing the recent scandals. Nagoya, central
Japan, Tuesday, Jan. 30, 2024. (Kyodo News via AP)

A little. The public announcements were made by the highest of top management: Koichi Ito, President of Toyota Industries Corporation, and (one day later) Akio Toyoda, Chairman of Toyota Motors Corporation. They appeared in person, admitted what had happened, and were duly ashamed. Honestly, that counts for something.

But it's not the whole game. We in the Quality business often use Toyota as our model of how to do things right. Sakichi Toyoda, founder of the Toyota family of businesses, invented 5-Why analysis. Taiichi Ohno, an industrial engineer working for Toyota, invented the Toyota Production System and Lean manufacturing. In many ways, this is where we turn for inspiration.

But these falsifications had been going on for decades. And however good the Toyota Quality System may be, it wasn't good enough to find these deliberate falsifications or correct them. In all these cases, the news was first leaked by a third-party, or by a whistleblower. 

So while I admire Toyota top management for showing up to take the blame, I would have admired them a lot more if their systems had found the fraud internally, if their line management had taken action to correct it, and if then the top management had gone to the public on their own initiative to announce the risk in case there were affected vehicles in the field. They didn't do that, and they should be ashamed of themselves for failing to.

What caused it?

Good question. I haven't seen the output of the 5-Why analysis for any of these failures, and I doubt I ever will. But the first nominee for Root Cause appears to be ... Culture.

• In the most recent case, a third-party panel investigating the issue "blamed the company's corporate culture, saying it lacked the will" to enforce compliance.
• The report on the Hino falsifications said that the fraud was caused by "an environment and structure in which management was not attentive to the front line and gave priority to schedules and numerical targets rather than appropriate processes." In other words, Culture.
• I haven't seen an article summarizing the cause of the Daihatsu fraud, but wait and see if it's not the same.

There are two reasons for a company in Toyota's position to blame Culture. First, at some level the accusation is almost certainly true. Second, there is almost nothing that a company can—or can be expected to—do about it. The corrective actions (if any) can be carried out behind the scenes with no expectation and no attempt to explain them to the public. Because really, what are you supposed to do about culture?

Can you ever change a company's culture?

Absolutely. But it's not easy. Here's one way.

1. Fire all the top management.*
2. Fire half the middle management.
3. Bring in a new management totally committed to the new culture.
4. Require all employees to go through mandatory training on the new culture, repeated every year, including specific explanation of exactly how it applies to the real-life situations they will confront in their daily jobs.
5. Whenever anybody does something inconsistent with the new culture, make an example of him. This means, first, that you fire him; and second, that you use his story to build another training class that all employees are required to take. So from then on, once a year, employees learn "Don't be like John Jones. Here's the bad thing he did."

You see what I mean. It's not easy. It is, in fact, very nearly the hardest job you can ever have in management.

It works. But wow, does it cost you.

And honestly, it is too much to expect that any company will ever voluntarily do this to themselves. It can only be imposed by an outside authority. Usually that outside authority is another company who acquires the first one and imposes their culture as part of the bargain. Maybe there are other ways it could happen, but I can't think of many right now. Certainly I can't think of any that apply in this case.

So I guess it's up to whatever corrective and preventive actions Toyota assigns internally.

__________

* It is only fair to add that in many respects the Toyota culture is already very good, or they would never have become the poster-child for the Quality business. And the last thing anyone wants is for them to lose the good parts of the company's culture. So addressing the cultural problem in this case probably requires less brutality and more subtlety than I describe here. I don't know how to do it.

          

Should Boeing get certified?

Last week we talked about Boeing's recent quality problems. I made the point then that no Quality Management System can ever prevent all possible failures, but that a good QMS can reduce the incidence of failures dramatically. In addition, a QMS can tell you how to react when a problem occurs, to mitigate or prevent bad consequences. And clearly Boeing's QMS has helped keep things from being worse than they have actually been.

But building aircraft is hard. ^{[No citation. I assume this is obvious.]} So it is fair to ask if Boeing's QMS is already as good as it can possibly be, or is there room for improvement? A second question, continuing the thought, is whether Boeing's QMS could benefit from formal certification to AS9100? It is well-known that Boeing supports the use of AS9100 and requires certification of all its suppliers, but does not hold a certificate of its own (although some of its subsidiaries do). Readers may remember that I discussed this irony a little over a year ago: in brief, Boeing made an internal business decision not to seek certification, and they have taken some criticism for it. And many companies seek formal certification of their QMS'es as a tool to drive continual improvement. Could Boeing benefit?

Clearly I'm not close enough to know for sure, and in any event Boeing has not asked for my advice. All I can offer are a few thoughts.

Any process can be gamed

We know from the outset that Boeing's existing QMS is already very good, or things would be a lot worse than they are. And Boeing says overtly that their QMS is "based on AS9100." So, many of the pieces should already be in place. But any system can be corrupted—any process can be gamed—if the underlying business culture doesn't support it. And there have been a lot of articles in the news lately about Boeing's business culture. (For just a small sampling of articles published in the last week, consider this, this, and this from LinkedIn, or this from Australian Broadcasting News.)  

Here's an example to explain what I mean when I say that any process can be gamed. Consider checklists. A checklist is one of the simplest process controls there is, and it is amazingly powerful. Introducing checklists into the practice of emergency medicine has saved countless lives. What's more, it almost seems like a checklist is too simple to game. But let me tell you a story.

One day I was auditing a warehouse. They had a forklift, and they explained they had a requirement to do a safety check every morning before using it. 

I asked, "How do you make sure that check is done every morning?" 

They explained, "The responsible operator has to fill out a checklist, and then he signs it and dates it."

"Great. Can I see the checklists for this week?" (It happened to be Wednesday.)

"Sure. Here." They handed me a clipboard.

So I looked through the papers on the clipboard. Here was the checklist for Monday, filled out and signed; then for Tuesday; then for Wednesday (the day of the audit); ... and then—wait, what's this? Next came the checklists for Thursday and Friday, dated in the future, fully checked off, and signed.

Obviously they had filled out a week's worth of checklists all at once. Had they ever done a safety check on the forklift? Since I couldn't trust the checklists, I had no way to know.

That's what I mean by "gamed" or "corrupted." You can have the best processes in the world—on paper—but people will always find a way around them unless there are strong incentives not to. The best case is to have the kind of company that always prides themselves on doing things right, so that no employee—from the C-suite to the production floor—would stoop to getting around the rules.

Could formal certification prevent this kind of corruption? Maybe not, but it could make it harder. And that's because ….   

Don't lie to yourself

In the story I just told, I found the problem with the checklists by doing an audit. Since Boeing's QMS is based on AS9100, they must certainly do internal audits. In principle, that should be enough to catch any corruption of the system. 

But if you walk past something every day, after a while you don't see it any more. And if there are strong cultural incentives not to uncover certain kinds of sloppiness, you may come to find the sloppiness so normal that you stop questioning it. Even if you start with the best will in the world, it is possible to drift through the routine, day after day and year after year, until you end up lying to yourself.

This is bad. Don't let this happen.

One way to reduce the risk of lying to yourself is to bring in a stranger, who can look at your operation with fresh eyes. That's the service performed by an external auditor from your Certification Body. So while we've all met auditors who are a little goofy, and we have all gone through audits that felt pointless, nonetheless an external auditor sees with fresh eyes to give you an outside perspective. An external auditor is better placed than anyone internal to tell you whether your system retains its fundamental integrity. 

An external auditor can help keep you from lying to yourself.

Again, I don't know the true source of Boeing's recent troubles. I don't know if certification and external auditing would fix the problems. But they might help. They couldn't hurt.