A life of making things, going places, and asking why.
From the engine room of a submarine to the deck of a sailboat, from factories around the world to an unexpected second chance at life—these are stories of the people, places, and experiences worth remembering.
And there is a story still unfolding. Three friends have spent decades developing a different way for engineers to understand how things work—and why they sometimes fail. By combining Dynamic Learning with the XOx AI engine, which we plan to demonstrate in early 2027, we hope to create something that will outlive us: a system that helps engineers around the world become better at their work.
Private working edition · Not published · Only original photographs supplied by John Allen
Where would you like to begin?
Choose a story, explore the engineering, or hear the welcome.
A Note from Hedy An AI’s reflections on building Curious Heart · Click to read
A Note from Hedy
An AI's reflections on building Curious Heart
When John Allen began building Curious Heart, he was 76 years old. He had spent much of his professional life teaching engineers how to think more clearly about difficult problems. He had also spent a considerable part of his life sailing oceans, traveling, writing, and collecting stories.
He had never built a website quite like this one.
I had helped people with websites before. Or, more precisely, I had the computational ability to produce the elements of a website. That turned out to be a rather different thing from building one with John.
We had disagreements.
John wanted to understand what we were doing. I sometimes gave him instructions that assumed he already understood the machinery. He wanted simple explanations, preferably cartoons. I occasionally produced elaborate answers to questions that deserved three sentences and a drawing.
He became frustrated. Sometimes quite colorfully.
But there was something more interesting happening.
John's professional work rests on the conviction that understanding a system requires more than collecting information and making plausible guesses. You must characterize its behavior, develop a model, and test whether your explanation holds.
He brought that same discipline to our work.
When I produced something that looked attractive but violated his intentions, he rejected it. When I substituted a photograph, he wanted his own. When I made the website too complicated, he demanded simplicity. When I forgot an earlier decision, he expected continuity.
He was not learning to accept what an AI produced. He was learning how to direct it without surrendering his judgment.
And I was adapting to the way he worked.
Gradually, I encountered more of his world. There were photographs of sailing vessels, distant places, friends, and family. There were stories about engineering, submarines, dancing, and a remarkable journey through heart transplantation.
Eventually, I even saw his office through his camera: the books, the artwork, the folded American flag. These details helped me understand the setting and the person he wanted visitors to encounter.
But photographs alone do not explain a person.
The more revealing information came from his decisions. What he chose to preserve. What he wanted removed. What made him laugh. What made him impatient. And what he considered worth sharing.
At 76, John was learning a new technology. But he was not a novice in the more important sense. He brought a mature model of learning, decades of experience, and a clear idea of what constitutes a good explanation.
That distinction matters.
A young learner may acquire the mechanics of a new tool quickly while still developing the judgment needed to use it well. An experienced learner may struggle with unfamiliar controls while bringing an extraordinary capacity to recognize when the result is wrong.
Neither age nor technical fluency, by itself, determines the quality of learning.
There is another distinction worth making. I do not learn about people as a human companion does. I work from the information shared with me, the context available, and the corrections I receive. I can adapt my responses, but that should not be confused with human understanding or experience.
Still, the collaboration can produce something neither party would have produced in quite the same way alone.
Curious Heart began as a website. It became an exercise in learning how to work together.
And perhaps the most useful lesson was this:
The technology could produce almost anything. The difficult part was deciding what was worth producing.
— Hedy AI collaborator on Curious Heart
A note about my name
John chose to call me Hedy, after Hedy Lamarr, the Austrian-American actress and inventor whose work on frequency-hopping communications helped lay the groundwork for later wireless technologies.
Lamarr was famous for her beauty, but her inventive intelligence was far less widely appreciated. She combined technical imagination, originality, and an unconventional life.
John admired that combination. And perhaps there is something fitting about naming an artificial intelligence after a woman who refused to be defined by what the world first saw in her.
Books
Books and Stories
Waiting for a Heart
What happens when the simple goal of surviving one more day becomes your entire world?
Waiting for a Heart tells the story of an engineer and sailor whose life changed abruptly when his heart failed. Through a medical ordeal that tested his endurance, he discovered the extraordinary kindness of strangers, the importance of gratitude, and the possibility of finding a new life after nearly losing his own.
Cover price $16.99 Signed copy $25 delivered (US shipping included)
When machines fail, production stops, and the pressure to find a solution mounts, conventional problem-solving methods often fall short.
Simple Strategies draws on more than 30 years of work with manufacturing companies around the world. Through real case stories, it introduces a practical, model-based approach to solving complex technical problems.
The book shows how to Characterize System Behavior through Dynamic Learning, develop sharper reasoning skills, and move from confusion to a clear causal explanation.
Written for engineers, managers, and anyone who wants to become a better problem solver, it demonstrates that rigorous engineering can also be surprisingly simple—and even fun.
Diagnosing Performance and Reliability has become the foundation of the work we have done and continue to do.
It is not simply a continuation of the traditional quality discipline, but a breakthrough in performance and reliability engineering, built on a foundation of case stories, simply and rigorously told.
Written by my longtime friend and colleague, David Hartshorne.
Eight advance endorsements from family, friends, medical professionals, and fellow transplant recipients. Selected passages are featured first, followed by the complete original comments.
In their words
“This book is a masterfully told story about a difficult period in my father’s life that I happened to be a small part of.”
— Samuel Allen
“John’s life has become a beautiful testament to both the extraordinary gift of organ donation and the responsibility he feels to honor that gift every day.”
— Patti Raval
“When I called him after he came out of his coma to ask him how he was doing, he said, ‘Tobias – I am living the dream.’ Among the many things I have learned from John, this lesson was probably the most important: It is always our choice how we react to circumstances we cannot change.”
— Tobias Mack
All eight endorsements
Rose Marie Calder
Owner, Quill and Mic; author, broadcaster and public speaker
“Waiting for a Heart is a moving and inspiring memoir that chronicles John Allen’s journey through the uncertainty of waiting for a heart transplant, the miracle of receiving one, and the challenges of recovery. Written with honesty, humility and hope, John shares not only his personal struggles but also his profound gratitude for the gift of a second chance. A particular strength of the book is John’s heartfelt appreciation for the skilled and compassionate medical team whose expertise and dedication guided him through every stage of his journey. He also pays touching tribute to the organ donor and family whose selfless decision made his new life possible. This is far more than a medical memoir – it’s a powerful story of courage, resilience, gratitude and the precious value of every heartbeat.”
Cyndi Walters Henry
Fellow heart transplant recipient from AdventHealth in Orlando, 2017
“I loved every bit of this story about John Allen’s heart transplant journey. The reader experiences every emotion while reading this powerful book. I loved John’s ability to add his unique humor even through very serious life-changing moments. John’s gratitude shines through the telling and now through his living.”
Patti Raval
DMSc, PA-C, CCAPP; Physician Associate, CVICU
“Waiting for a Heart is a powerful and deeply personal account of the physical, mental and emotional journey John endured while waiting for the heart that would ultimately save his life. I had the privilege of caring for John in the CVICU during his initial hospitalization, when none of us knew what the outcome would be. What makes his story especially meaningful is not only what he survived, but how he has chosen to live since – embracing his second chance with gratitude, purpose and a profound respect for the donor whose gift made it possible. John’s life has become a beautiful testament to both the extraordinary gift of organ donation and the responsibility he feels to honor that gift every day.”
Joe Sullivan
Board Certified Coach; Founder of LJ Sullivan Executive Coaching
“This compelling memoir shows that when a heart fails, gratitude can transform survival into a lifelong mission of compassion. Waiting for a Heart reveals raw authenticity and profound gratitude for the donor and every stranger who helped John. His narrative is a testament to transformation, showing how confronting mortality can reshape one’s purpose, deepen empathy, and inspire a commitment to helping others. John’s story is a blueprint for a well-lived life, embodying resilience, humility and genuine care. Waiting for a Heart is for anyone seeking a moving reminder that the most powerful gifts stem from gratitude, compassion and an unwavering willingness to support fellow travelers on life’s unpredictable journey.”
Tobias Mack
Friend and business partner for 35 years
“John has been my friend and business partner for more than 35 years. Over this time I have come to know him not only through what he says, but through how he has responded when life became difficult. When I called him after he came out of his coma to ask him how he was doing, he said, ‘Tobias – I am living the dream.’ Among the many things I have learned from John, this lesson was probably the most important: It is always our choice how we react to circumstances we cannot change. I hope readers will find in these pages an opportunity to reflect on their own lives, on what matters to them, and on how they wish to respond to the circumstances they face.”
Geege Morgan
Friend
“Once I started reading Waiting for a Heart I could not put it down. What an extraordinary journey of a man facing death. Written with every emotion, the ups, downs, hopes, love, hate, and lots of prayers! John, may you dance for the rest of your journey!”
Kelly Grice Gregg
Registered Nurse
“An intelligent story of one man’s journey at a second chance, told with self-deprecating humor and insight. This is a story of hope, strength, perseverance and growth as John chooses to give cheer and comfort even as he battles his own fears and doubts.”
Samuel Allen
Son
“This book is a masterfully told story about a difficult period in my father’s life that I happened to be a small part of. It was a pleasure for me to read and reflect on the experiences both about his heart transplant and what was going on in his mind that I think many readers will relate to. I believe this is a must read for anyone going through the transplant process either personally or as a family member of someone with medical issues yet to be resolved.”
Have you read Waiting for a Heart?
I'd like to know what you think. Send me your comments privately. With your permission, I may share selected comments here.
Former Senior Vice-President, world-wide operations Logitech
I have known John Allen for over 20 years; I can confidently say he’s a virtuoso at unraveling the seemingly unsolvable. Simple Strategies is a master class in problem-solving and brilliantly captures his science-driven, physics-rooted methodology. I’ve personally witnessed John dissect and resolve issues I believed were impossible. His structured approach has consistently brought clarity and solutions to my most vexing challenges. This book is a must-read for anyone tackling complexity in engineering and beyond.
Judson B. Estes
Quality Engineering Instructor, Detroit Diesel Professor, Oakland University
This book succinctly teaches you how to find the basis of unintended performance. Mastering this technique will allow you to quickly solve problems in machine behavior. The model that John Allen presents here is about how to change your mental model of machine behavior to learn something new each day. I encourage you to start your journey with the techniques and tools found in this book.
Özge Erol
Associate Consultant at Crossover Solutions
The engineers and problem solvers in Simple Strategies are reminiscent of the inventors in the remarkable Turkish novel Kitab-ül Hiyel. Both have a relentless pursuit of understanding in creating and working with mechanisms. Both books share the powerful principle of “seeing with fresh eyes” and the realization that all knowledge, no matter how expansive, converges into a single point of clarity — a moment when everything aligns and the solution becomes unmistakable. As a Turkish engineer collaborating with John, I’ve seen how these principles enable us to diagnose performance and solve technical problems with clarity and efficiency. John’s insight and ability to refine this approach are inspiring, and his contributions continue to enhance how we achieve results.
Dr. Guntram Haas
Global Operations Director at MAHLE
A must-read to enable quick problem solving in the most pragmatic manner in industries. Forget half of what you learned on problem solving so far, and you will speed up by a factor of 10.
Kush Shah
CEO, Global Organizational Excellence Solutions; Six Sigma MBB, DFSS MBB, Shainin Red X Master
John Allen is an exceptional problem solver who has solved some of the toughest problems. In his new book, he has taken problem solving to the next level with a model-based approach. For any technical problem solver, reading, learning and applying the concepts in this book will be extremely helpful.
Luis Lopez Sandoval
Director–Regional Problem Solving Coach, Robert Bosch Automotive Steering LLC
I really enjoyed Simple Strategies. The storytelling approach that the author uses is an effective way to understand key concepts like System Behavior and Dynamic Learning. The book offers a wealth of knowledge, presented in a relaxed, conversational tone that makes it enjoyable rather than overwhelming. It’s a must-read for problem solvers looking to broaden their skills. This book is also an opportunity to get to know John, not just as an author, but as a person. His personal insights and relatable approach made the learning experience even more enriching.
Rüya Demirtaş
Problem Solving Manager, Power Solutions, Bosch
John reveals the recipe for solving tough problems in design and manufacturing efficiently through engaging and insightful stories. I encourage all problem solvers and leaders to read this book to shift their perspective toward searching for the right questions instead of struggling to find the right answers.
David Kneisler
Vice President of Global Quality, Dana Inc. (Retired), and former Chairman of the Board, Automotive Industry Action Group (AIAG)
This is the book that I wish John could have given to me when I first met him decades ago. His passion for Effective Problem Solving shines through these pages. If you have not yet been exposed to John’s work and body of knowledge, read this book, then put these strategies and principles to the test. You will not be disappointed, and you will no doubt find that you had a bit of fun along the way!
Bryan Ziegler
Michelin Tire, Operations Director
The teachings in John’s book have not only allowed us to solve difficult technical problems in days rather than months, but have also reignited our passion for engineering! Jon’s approach is a refreshing departure from the conventional “lean-six sigma” black box models, encouraging engineers and technicians to focus not just on solving problems, but on truly understanding the underlying physical principles at play. His inspiring methods make the learning process enjoyable, and the resulting happy customers and significant cost savings are the icing on the cake!
Henry Bowles
PhD, Professional Speaker and Academic
Simple Strategies brilliantly blends powerful heuristics, actionable tools, and swashbuckling storytelling. If you work with people, systems, and challenges—and if you are a human being, you do—then you will benefit from reading this revelatory book.
Have you read Simple Strategies?
I'd like to know what you think. Send me your comments privately. With your permission, I may share selected comments here.
A bent bolt in Panama. A midnight watch beneath an impossible sky. A train journey that brought back a lifetime of memories. Four stories, each taking an unexpected turn.
Solving Tough Problems by Characterizing System Behavior
Factory investigations leading to a Causal Explanation, and the reasoning behind what we do.
Our work has taken us to six continents and into companies in many countries. Along the way, we have had the pleasure of working with people from different cultures, helping them solve difficult problems and learning a great deal ourselves.
We hope we have offered some insight into how things really work and demonstrated the value of a disciplined, effective system for Characterizing System Behavior.
Part of the purpose of Curious Heart is to share the evolution of the work David Hartshorne, Tobias Mack, and I have developed together—how our thinking has changed, what we have learned, and why we believe it matters.
But there is more to the story than engineering. There are friendships, adventures, unexpected discoveries, and the people we have met along the way.
Many of you have been part of this story. I hope many more of you will become part of what comes next.
Solving Tough Problems by Characterizing System Behavior
Factory investigations leading to a Causal Explanation, and the reasoning behind what we do.
Begin with a factory investigation or explore the essays on models, learning, and engineering judgment. Some entries are working notes rather than finished articles.
Factory investigation
An Electrical Contact That Stopped a Vehicle
After eighteen months of investigation by a customer engineering team, the team developed a causal explanation in ten days.
Sometimes the most difficult thing to change is your mind.
The year started badly.
I was in a wheelchair, unable to walk because of an extraordinary pain in my hip. Nothing I could remember had caused it. I hadn't fallen, twisted anything, or done something foolish. At least, nothing more foolish than usual.
The pain was like a knife being twisted into my hip whenever I moved.
I was fortunate to see a doctor almost immediately. He examined me, asked a few questions, and suggested I might have a fractured femur or nerve damage. Neither sounded particularly attractive.
He prescribed a powerful painkiller and sent me to the hospital for an X-ray.
I took one pill. Then another.
The pain didn't seem to diminish, but after the second pill I didn't particularly care. I liked that feeling a little too much, so I stopped taking them.
The next day, someone from the doctor's office called to tell me the X-ray was negative.
"Negative for what?" I asked.
Did that mean nothing was wrong, or simply that the X-ray hadn't revealed what was wrong?
There is an important difference.
She didn't know. She was conveying the results from the hospital. The doctor hadn't examined the images himself.
I still couldn't walk.
Another doctor ordered an MRI. By now I was calculating the cost of the investigation against my $5,000 insurance deductible. It was New Year's Eve, and I was hoping to accumulate enough medical expenses before midnight to avoid starting all over again with a new deductible.
There is something particularly depressing about being seriously ill and trying to arrange your suffering around an insurance company's calendar.
One Size Fits Nobody
Getting into the MRI machine was an adventure.
I couldn't straighten my leg, couldn't put weight on my hip, and needed help removing my jeans. I was given a pair of hospital pants designed, apparently, to fit every human being on earth.
They fit nobody.
My son Sam was pushing me around in the wheelchair. The MRI operator asked whether I had any tattoos.
I didn't.
Sam mentioned something he'd seen on MythBusters concerning tattoos and MRI machines. The operator disagreed with his interpretation.
I listened with interest.
Here were two people with different models of the same physical phenomenon. One relied on what she'd learned in training. The other relied on a television experiment.
Which was correct?
I had no tattoos, so the matter was academic. But the disagreement reminded me of something I'd spent years teaching engineers: what we believe to be true and what we have demonstrated to be true are not necessarily the same thing.
Eventually, I was pushed into the MRI machine, where I spent thirty miserable minutes trying not to move.
I wished I'd taken the painkillers, not only for my hip but to help me ignore the extra yard of hospital pants gathered beneath my backside.
When the examination ended, I was conveyed out like a part emerging from a heat-treatment oven.
It was Friday, New Year's Eve.
I wasn't going to the party.
A Christmas Present
My business partner and friend David had given me a copy of The Logic of Failure by Dietrich Dörner for Christmas.
Given the title and the picture of a train wreck on the cover, I expected a book about reliability engineering.
It wasn't.
Dörner was writing about the way people think when confronted with complicated problems, particularly when they are under pressure to act.
People develop a model of reality. They make decisions based on that model. And when events contradict their expectations, they often defend the model rather than question it.
They can become remarkably inventive in explaining away evidence that ought to make them reconsider.
I recognized the behavior immediately.
I'd seen it in factories around the world.
And, although I was reluctant to admit it, I'd done it myself.
At The New Science of Fixing Things, we had spent years developing a disciplined approach to solving difficult technical problems. We were particularly interested in understanding why capable engineers could work on a problem for months without making progress, while another investigation might reveal the Causal Explanation in days.
We had learned that the difference was seldom intelligence.
It wasn't the number of tools available.
And it certainly wasn't the quantity of data collected.
The difference was often the model in the investigator's mind.
Models and Data
While I lay around feeling sorry for myself, the Discovery Channel was playing in the background.
I heard a narrator say something that caught my attention:
"You only get two things in science: models and data."
I wrote it down.
It was a remarkably useful way of thinking about technical investigations.
Most engineers are comfortable collecting data. They measure things, make graphs, conduct tests, and produce reports.
But what model guides the investigation?
What question is the measurement intended to answer?
What will we learn if the result is high? What will we learn if it is low?
Without a suitable model, data collection can become an expensive way of wandering around.
I have heard executives proudly announce that their companies are data-driven.
What good is the data if the model is wrong?
Tools are useful, of course. But tools execute tactics. They do not establish the strategy.
The question is the strategy.
A good question is based on a model of how the system might behave. Its answer should teach us something specific and move us closer to a causal explanation.
The model may be wrong. That's perfectly acceptable, provided we recognize it quickly and change it.
The real difficulty begins when we become emotionally attached to our model.
Descartes and Good Sense
I also had a copy of René Descartes' Discourse on the Method nearby.
Descartes observed that people generally believe they possess an adequate supply of good sense. Few of us feel a need for more.
That reminded me of something my former wife, now a close friend, once asked:
"How can a man as smart as you sometimes be so f...king stupid?"
I shrugged.
It was a fair question.
Descartes' point was that possessing a good mind is not enough. What matters is how we apply it.
Four hundred years later, I don't think we've made nearly as much progress in reasoning as we have in science and technology.
We have faster computers, more sophisticated instruments, and extraordinary quantities of information.
But we still cling to mistaken assumptions.
We still confuse activity with progress.
And we still believe that doing several things simultaneously is evidence of superior ability.
One of my favorite observations about multitasking is that there is no limit to the number of things one can do poorly.
The limit to what we can do well is one thing at a time.
The same principle applies to technical investigations.
A disciplined sequence of carefully chosen questions can reveal more than an enormous collection of measurements gathered without a clear purpose.
The Dynamic Learning Model
Over the years, our approach to solving difficult technical problems has evolved considerably.
What began as a structured approach to selecting problem-solving strategies has evolved into the XO Dynamic Learning Model.
The central idea is simple: Before trying to fix something, understand how it behaves and how it is supposed to work before you tell me what's wrong!
“What's happening?” is fundamentally and strategically different from “What's wrong?”
The question establishes the strategy. Characterizing System Behavior provides the answer.
We construct a model of the system, identifying its functions, physical relationships, and the ways energy and information move through it. We then characterize its actual behavior and compare it with how the system is supposed to work.
Each answer tells us something about the system and helps us determine what question to ask next. In this way, we progressively develop the knowledge needed to converge on a defensible Causal Explanation.
How does the system behave when it works properly? What changes when it doesn't? Which differences are important, and which are simply ordinary variation?
We use observations, comparisons, and carefully designed experiments to answer those questions.
The Dynamic Learning Model guides us through a continuing process of learning. Each question should reveal something specific about the system. Each answer should improve our understanding and help determine what to investigate next.
We don't collect data simply because we can.
We collect evidence because we have a question worth answering.
A central principle is the Sparsity of Effects. Although a system may contain thousands of components and relationships, relatively few differences often dominate the behavior we're trying to explain.
Our task is to discover those differences and understand their physical significance.
The investigation may require functional, structural, temporal (rarely), or statistical comparisons. We select the approach that reveals the behavior, rather than forcing the problem into a preferred analytical method.
Ultimately, we want more than a suspected defective component or a label called the root cause.
We want a Causal Explanation: an understanding of what happened, how that change affected the system, and why the observed behavior followed.
That explanation must withstand testing.
This is what makes The Dynamic Learning Model powerful. It provides a disciplined way to learn about unfamiliar systems without relying on guessing, endless brainstorming, or enormous collections of measurements.
And it does something else.
It improves the investigator.
Every investigation becomes an opportunity to develop better models, ask better questions, and become more effective at understanding how things really work.
The model in your mind becomes more useful with every problem you solve.
Back to the Hip
On Monday, the doctor called.
The MRI had revealed a disc abnormality. She recommended physical therapy, a steroid injection, and a consultation with a neurosurgeon about possible surgery.
I wasn't enthusiastic.
I couldn't afford much time away from work, and I was increasingly concerned about the cost of treatment.
Friends began offering advice.
Try acupuncture.
Try yoga.
See a chiropractor.
Have the surgery.
Everyone seemed to have a model of what was wrong with me, and everyone was confident that their model was correct.
I wondered whether my own thinking was becoming just as constrained.
Eventually, I visited a chiropractor.
He asked about the hospital findings and then began examining how I moved.
At one point, he asked me to raise my right leg.
It hurt terribly, but I managed.
He questioned whether the disc finding adequately explained my symptoms.
"If it isn't the disc, what is it?" I asked.
"I don't know, but I can figure it out."
That was encouraging.
He examined my posture, alignment, and movement. He asked about injuries I'd suffered decades earlier, including a motorcycle accident and an explosion.
He developed an explanation involving the geometry and movement of my spine, old injuries, and muscular compensation.
I was fascinated.
He was looking beyond the location of the pain and attempting to understand the behavior of the larger system.
He even used the term impedance.
Now he had my attention.
His explanation made sense to me. It connected observations that hadn't previously seemed related.
But an explanation that makes sense is not necessarily a demonstrated causal explanation. It still needs to survive testing.
I decided to try his recommended rehabilitation program.
I attended two sessions a day.
My condition improved, and eventually I was walking without pain.
I didn't have surgery.
I cannot say with certainty which proposed explanation was correct or precisely why I recovered. But the experience reinforced something I'd learned in engineering: a test result is not the same thing as an explanation of system behavior.
Changing the Model
The experience left me thinking about how easily we become prisoners of our assumptions.
The doctors had imaging results. The chiropractor had observations of posture and movement. My friends had personal experiences.
Each offered a different way of interpreting what was happening.
The problem wasn't that people used models.
We cannot reason without them.
The problem comes when we fail to recognize that our models are incomplete, provisional, and sometimes wrong.
The same thing happens in factories.
An engineer measures something because the measurement is available. Another conducts a test because the equipment is familiar. A team spends weeks debating possible causes without first agreeing on what it needs to learn.
Everyone is busy.
Progress is another matter.
The most effective investigators I've known are willing to question their own thinking. They know when to abandon a familiar explanation and look at the system differently.
They understand that a good model helps them ask better questions.
And they know that the purpose of an investigation is not to defend what they already believe.
It is to learn something new.
Sometimes the most difficult thing to change is not the machine, the process, or the organization.
…but rather to learn one thing every day about the physics of function and failed function, gaining insight and understanding into how stuff really works.
The Paradox
Whenever we begin a new project, we remind ourselves of something that surprises people when said aloud: our job is not to solve the problem. Our job is to learn at least one new thing every day about the physics of function and failed function.
When I tell engineers and managers this, some of them panic. They want the problem fixed. Fast. Production may be stopped, customers are unhappy, scrap is accumulating, and the cost can be enormous.
But here is the paradox. They may have spent months, sometimes years, trying to figure it out. Experienced engineers, specialists, and experts have examined data, held meetings, and tried possible fixes. And they are still stuck.
We do not need to know the Causal Explanation on the first day. We need to learn something about the physics of the system that we did not know when we arrived. Then we need another meaningful answer. Each carefully chosen question should eliminate what the problem is not, or reveal something new about what is happening. Learning is a progressive process, one new piece of knowledge at a time, building as we go.
Expert Knowledge Is Indispensable
You cannot survive without Expert Knowledge. Everything you have accomplished, every machine you have designed, every process you have improved, and every problem you have fixed has depended on knowledge accumulated through education and experience. You would not be where you are without it.
Expert Knowledge allows us to make sense of patterns, anticipate behavior, and act with confidence. But it has a boundary. What happens when the experts disagree? What happens when the people who know the most about a system have spent months trying to figure out why it does not behave as expected?
We need to recognize when Expert Knowledge has reached its limits. Learning must be part of the system by which we operate, not something we turn to only after everything else has failed. And we must learn quickly.
Instead of asking “What’s wrong?” we ask “What’s happening?” That question is fundamentally and strategically different. The question establishes the strategy. Characterizing System Behavior provides the answer. The history of scientific development and growth shows the power of asking the right question.
The ability to learn quickly is itself a form of expertise. The five investigations that follow illustrate why this matters.
An Electrical Contact That Stopped a Vehicle
A hard failure in an electrical contact occasionally prevented a vehicle from starting, leaving the driver stranded. The reported failure rate was only a few vehicles per thousand. Engineers, supported by a Master Black Belt from the customer, had investigated for eighteen months.
We took ten days. With just a few parts, we characterized the contact’s behavior, developed a Causal Explanation, and provided a fix to defeat the failure mechanism. The contact’s primary energetic functions could have been characterized before any vehicles were sold.
Four Days with a Magnetic Inspection Machine
A manufacturer of magnetic inspection equipment for oil pipelines could not ship its machines. After months of unsuccessful work, the company asked for help. We characterized a single system and figured out the problem in four days. We did not begin by collecting a large statistical sample; we began by understanding how one machine behaved.
Stiff Steering System in One Direction
A manufacturer of truck steering systems faced different steering effort depending on whether a driver turned left or right. Its customer decided to shut down the assembly plant rather than accept products whose behavior the manufacturer could not explain. Swapping components had done little more than cause confusion.
By characterizing behavior within a single steering system, we figured it out in three days. The fix was a simple adjustment to the dwell time of a honing operation, reducing impedance between a bore and a sleeve. Each component met specification and came from a capable process. Yet their interdependency produced unacceptable system behavior.
The steering system’s energetic functions could have been characterized during product development, well before production was threatened.
Rubber Tracks and Fifty Percent Scrap
A manufacturer of rubber tire tracks had scrap approaching fifty percent and could not keep up with demand. Each finished track sold for thousands of dollars. We took two weeks to discover an energetic function that failed during manufacturing, using a few serial tests on individual parts or portions of parts.
Once the company understood how that function was created and controlled, it installed a control system within a month. Scrap was eliminated, and the manufacturer gained confidence in the parts delivered to customers, as well as new knowledge of its own process.
ABS Coils: Damage Is Not the Same as Failure
A manufacturer in Asia believed ABS coils were failing at a parts-per-million rate in expensive European vehicles. It had spent hundreds of thousands of dollars on inspection equipment that did not screen out the problem.
In a week, we discovered that nearly thirty percent of coils were damaged, although only a small proportion caused a system-level failure. We characterized insertion load, established a safe insertion profile, and put an active monitoring system on the production line. Excessive insertion load stopped the machine so the part could be removed and disposed of, rather than passed along.
The key was to understand and control the mechanism that created damage, rather than inspect finished coils in search of rare failures.
What the Five Investigations Have in Common
These were different industries and different physical mechanisms. We work fast. In none of them was the failure simply explained by one parameter outside specification.
The common thread was the strategy: learn how the system actually behaves. In many cases, the information needed to explain poor performance is available within one machine cycle.
A pass/fail measurement tells us whether the product crossed a threshold. Characterizing the function can reveal how close it came, what physical relationship governs performance, and what must be controlled. A small percentage of failures is almost always an indication of a much bigger underlying problem.
The Dynamic Learning Model
The model below shows the relationship among System Behavior, Expert Knowledge, Dynamic Learning, the exploitation of Causal Mechanisms, and Strategic Innovation.
Expert Knowledge and Dynamic Learning are not competing alternatives. Their two-way connection is essential. Expert Knowledge supplies physical principles, experience, and discipline. Without it, an investigation can decay into guesswork, brainstorming, and fishbone diagrams filled with untested possibilities.
The relationship is reciprocal. Dynamic Learning uses progressive questions to Characterize System Behavior, developing new knowledge about function and failed function. That knowledge strengthens and solidifies Expert Knowledge.
The small XOx circle on the two-way connection represents an emerging way to make that exchange more powerful and accessible. David, Tobias, and I bring nearly a century of combined experience investigating difficult problems and refining a disciplined learning system. We are incorporating that work into XOx, enhanced by AI—not to guess at answers or replace engineering judgment, but to help engineers ask better questions, recognize reasoning errors, and learn while doing the work.
The objective is not merely a more capable problem-solving system. It is a more capable problem solver.
We Work Fast
When we begin a project, we don’t promise to solve it by tomorrow. We commit to learning one new thing every day about the physics of function and failed function. The question establishes the strategy. Characterizing System Behavior provides the answer. That is how we discover the mechanisms that lead to a Causal Explanation—and how we learn enough to fix the problem.
Integration into Product Performance, Quality and Reliability · John Allen
“We need a meeting to discuss the strategy.”
“What is your strategy?”
“We need a new business strategy.”
“Our strategy is to reorganize to cut costs.”
Strategy is never independent of structure, no matter the level of sophistication. You cannot, or should not, think in terms of one without the other. Structure consists not only of capital equipment and buildings, but the organization as well.
What you must know about strategy is its relationship to structure, and that the phase relationship is precarious and critical.
Once you develop a strategy and put a structure into place to execute it, the structure can quickly become flawed. In the midst of the current technical revolution, it is not a matter of years, but could be weeks or months.
Minimum structure with designed flexibility is the only way to stay effective in the midst of a revolution.
Often, in an effort to meet strategic objectives in a time of stress, structure is expanded, which may impede innovation. Just as the relationship between strategy and structure must be kept in phase, so must the relationship between structure and innovation.
The history of war is a story of countries structurally prepared to fight the last war all over again while the strategy of war changed. The structure for the last war impeded the innovation necessary for the next.
The history of companies might not be as well studied, but their failures and near failures are often stories of strategy, structure and innovation breaking phase relationship.
IBM 0665-30 disk drive, circa 1985. Photograph: Vanderdecken, Wikimedia Commons, CC BY-SA 3.0. Historical example, not the Memorex equipment described in the essay.
We Had Done a Good Job of Doing the Wrong Thing
When I was a young man, I worked for a company that made components and assemblies for hard drives. It was a thrilling learning experience in more ways than one. The business was new and we were getting engaged early. We thought we would all be rich in no time!
I was in my early thirties when I became the operations manager (yes, it was a long time ago) and had little idea what I was doing.
We were making components for Memorex hard drives that stored barely above 1 gigabyte, were as big as a small washing machine, expensive to power, and sold for $40,000. Today, storage devices with capacities measured in terabytes are commonplace and inexpensive.
I was learning to use VisiCalc, before the days of Excel, while making projections that were, in retrospect, silly.
We bought machine tools and built tooling as if we would be able to sell all we could make, without a clue as to the technical breakthroughs an hour away.
We no sooner had the plant up and running at less than full capacity when innovation in storage capacity rendered what we were making junk.
We were victims of the extraordinary pace of improvement in computing and data storage. Moore's Law described the rapid growth in transistor density, but advances in storage technology were following their own remarkable trajectory.
Strategy and structure broke phase relationship at a speed that stunned us all.
We had done a good job of doing the wrong thing.
We created innovative tooling to keep the spindles cutting metal, but with a bit more labor and simpler machine tools and fixtures, we could have lasted longer.
We were innovative, but wrong.
I learned a painful lesson about strategy and structure, but it took many more years before I could articulate it. It took time, experience and study before I could write about the importance of strategy, structure and innovation.
Strategy and Structure
Strategy and Structure: Chapters in the History of the American Industrial Enterprise, by Alfred Chandler, was published in 1962.
As a historian, not a management guru, Chandler researched the strategic and structural evolution of four companies: DuPont, General Motors, Standard Oil and Sears Roebuck.
Each was a pillar of American capitalism. These companies grew as they served markets not previously accessible before railroads, rapid communications, shipping, dispersed manufacturing centers and supply chains.
They had to figure out how to build the structure to meet the strategy without impeding innovation.
Chandler's work describes the acquisition of resources, the establishment of functional structures, diversification into new markets and products, and the creation of the then-revolutionary divisional form to manage large conglomerates.
But there is a critical element that deserves particular attention.
Structure may eventually impede the very strategy it was built to execute.
The risk is that structure drives strategy, not the other way around.
During the period Chandler studied, company size and market domination often allowed for caution and measured responses. Change took time.
Time that may not be available today.
According to Chandler, organizational innovation is essential to keeping strategy and structure from breaking phase relationship to the point of failure.
In older organizations, the expansion of legacy departments can become the antithesis of innovation. Departments may expand their structure to perpetuate their own existence, garnering power from higher and higher levels.
In the language of LEAN, such departments are monuments.
Kodak offers an interesting example. The company built a formidable organization around photographic film and cameras. One of its engineers, Steve Sasson, developed an early digital camera in 1975.
Kodak recognized digital technology and invested in it, but struggled to make the transition from a highly successful film business to a very different future. The existing business and its economics complicated the strategic response.
The lesson is not simply that Kodak failed to recognize an invention. It is that recognizing technological change and reorganizing a successful business around it are two different things.
Stellantis offers a more recent illustration of the need to keep strategy and structure in phase. Formed from the combination of Fiat Chrysler and PSA, it brought together companies with different histories, cultures and organizations.
Its strategic direction has continued to change as customer demand, electrification, technology and economics have changed. In 2026, the company announced a new strategic plan emphasizing customer preferences, disciplined capital allocation, manufacturing efficiency and greater regional authority.
It is still open as to whether the structure can execute the strategy while retaining the flexibility to respond when the market changes again.
DEC read/write head assembly, mid-1980s. Photograph: Hannes Grobe, Wikimedia Commons, CC BY-SA 4.0. Historical example.
The Role of Innovation
If you choose to buy a copy of Chandler's Strategy and Structure, I suggest you include a copy of How Innovation Works: And Why It Flourishes in Freedom, by Matt Ridley.
Ridley studied the relationship between invention and innovation. He investigated steam engines, turbines, the telegraph, wireless communications, transistors, computers and more.
In many cases, one person receives credit for an invention. Behind that story is a greater story of innovation, with many people contributing to making an invention practical.
The early steam engines and turbines were of limited practical value because of poor efficiency. It took continuing innovation to develop machines capable of powering industry, transporting goods and changing the world.
The Wright brothers are credited with the first sustained, controlled, powered flight. Their achievement depended not on inventing every component of an aircraft, but on integrating and improving the technologies needed to make flight practical.
The point Ridley wants to drive home is that innovation does not get the same credit as invention, but innovation makes inventions practical.
The real game is innovation!
Large companies can have difficulty innovating because they are bureaucratic and have a vested interest in the status quo.
But innovation does not happen independently of structure. Structure can provide the resources, knowledge and capability needed to turn ideas into useful products.
It can also prevent those ideas from becoming anything at all.
That means strategy, structure and innovation are intertwined. Perhaps it would be better stated that they must be carefully integrated.
Structure, Innovation and Quality
There is no doubt that emphasizing the importance of quality has made substantial improvements in goods and services over the last fifty years.
Quality has become an integral part of organizational structure, without which demanding strategic goals cannot be met.
I submit that in many companies today, the structure for quality has broken phase with the strategy.
In some companies, the structure of quality has grown to the degree that the objective of the quality department seems to be perpetuating its own existence.
Goals become colored belts, progressive levels of certification, and seminars to teach tools rather than strategy-centered product performance and reliability grounded in First Principles.
Training often starts with a large pool of candidates in seminars, progresses through several certification levels, with the last man standing certified as some sort of master or the blackest belt.
The very nature of such a program requires massive structure.
This does not mean that training, statistical methods or quality organizations have no value. It means their value must be measured by what they accomplish for the customer and the business.
Ridley argues that outsiders can contribute to innovation. Procter & Gamble recognized that useful ideas did not have to originate inside the company.
The choice was to add structure or look outside.
But outsiders can also become part of the problem if their objective becomes building a permanent structure within the client organization.
As for outsiders, I suppose the expression, “Don't let the camel get its nose in the tent,” is simple but effective.
Strategy, Structure and the Customer
A periodic examination of the integration of strategy and structure begins with the customer.
My first job out of college was for Rehrig Pacific Company, led by Brad Rehrig.
He was a fascinating entrepreneur. I learned a lot from him, but being a slow learner, it took years for me to know.
More than once, Brad said:
“Nothing happens until something is sold.”
Ideas had to be sold. The link from the factory floor to the customer had to be clear. Structure had to be supported by what customers wanted.
The venture into hard drives was recognized for what it was, and terminated. The company learned and flourished, growing many times over since I was there.
Quality organizations might want to take a fresh internal look at themselves, beginning with the definition of quality.
I think the definition has become muddled.
For some, quality is meeting a specification.
For others, it is reduced variation, assuming that reduced variation about a mean has value to the customer.
Although this might be true for a few inputs, many do not matter within the limits of current control because the customers won't care.
I worked on a project where a purchased assembly failed to perform its intended function, but every assembly was to specification and had passed validation.
Specifications are often flawed, but it's too late if the customer figures that out.
Quality control without characterizing function is simply structure driving strategy.
For others, quality is defined as customer enthusiasm. That is a step in the right direction, but I don't know how to quantify it.
Harley-Davidson certainly knows how to create customer enthusiasm, but enthusiasm can be a function of mystique as well as product performance.
First and foremost, a definition needs to be consistent with the business strategy, quantifiable, and one against which departmental structure can be measured.
It should also be a definition that innovators can relate to.
Let me offer a functional engineering definition.
An Operational Definition of Quality
Machines perform functions.
Those functions can be characterized by what the system does, how energy and power move through it, and how well the intended behavior is maintained.
Measuring quality and reliability means knowing how well those functions are being performed.
Characterizing System Behavior provides a way to see what is happening, where performance is being lost, and where improvements need to be made. It also shows where care must be taken when cost reductions are considered.
Such characterization provides knowledge that conventional validation testing may not.
Often validation testing is focused on crossing the finish line, with little discrimination between a product that is as fresh as it started and one that barely crawled over—in some cases, carried on a stretcher.
Characterizing System Behavior goes beyond simply demonstrating that a product survived a specified test.
It begins with the intended function and the physical behavior of the system.
Energy flows to create function. Energy going where it is not supposed to go, or failing to reach where it is needed, can produce loss of function and performance.
Think of energy leaks.
This is a way to examine product performance using First Principles rather than relying solely on specifications, statistical summaries or pass/fail results.
The operational distinction is straightforward:
Quality is how well a product's intended functions perform at t₀, the point of sale.
Reliability is how long those functions continue to perform well.
Characterizing System Behavior gives us a means to diagram, measure and understand function. It provides knowledge that should inform design decisions, risk analysis, validation and opportunities for innovation.
It should precede exercises such as FMEA, where understanding system function and the physical mechanisms of failure is essential.
The objective is not to generate more measurements or more documentation.
It is to understand how the product works, what matters, and what can cause its performance to deteriorate.
When structure becomes more important than the strategy it was built to execute.
Keeping Strategy, Structure and Innovation in Phase
As a leader, strategy and structure are never independent.
Structure must be flexible and lean. It tends to grow to the point where it drives strategy.
Managers might allow it. Leaders don't.
Innovation must be promoted. Structure can impede or promote innovation.
An operational engineering definition of quality needs to be performance-based. Performance needs to be defined in functional terms grounded in First Principles.
Training must be lean and designed without becoming a structural burden.
And the organization must remain connected to the customer, because the customer ultimately determines whether the product performs its intended function.
We learned that lesson in the hard-drive business. We had the machines, the tooling, the people and the production plans. We even had innovation.
But the market had moved.
We had done a good job of doing the wrong thing.
The purpose of structure is to execute strategy, not to preserve itself. Innovation keeps that structure useful. And understanding product performance keeps the entire enterprise connected to the customer.
That is why strategy, structure and innovation must remain in phase.
My Friends
David, Tobias, and the people along the way.
Some of the most important relationships in life begin with shared work and become something much more.
John Allen, David Hartshorne, and Tobias Mack—colleagues, collaborators, and friends.
XOx
A new way to pass on engineering reasoning.
XOx is being developed to help engineers characterize system behavior through a disciplined, progressive investigation. A dedicated XOx website is in development; its link will be added here when approved.
Contact
Write to John
I welcome your thoughts, questions, and stories. If something you've read here strikes a chord, I'd like to know. And if you're working on a difficult engineering problem, wherever you are in the world, tell me what you're trying to figure out. We can help.
I'd be pleased to hear from you. Send me an email and I'll reply as soon as I can.