The Process Bigot
Process and the Secrets to Quality
A few days ago, I wrote a piece about the day I learned what Quality actually is, and how it changed my life: erictelltales.substack.com/p/smashed-to-splinters-designed-for. It tells the story of how I came to understand Robert Pirsig’s Zen and the Art of Motorcycle Maintenance, and how applying his theory changed my life.
Quick recap, if you haven’t read it: I was trying to reassemble a bed I’d built two years earlier out of cheap packing-case wood. The screw holes had worn loose from being reassembled twice already, and after three hours of clamps, glue, and collapsing legs, I lost my temper completely — the only time in my life I ever have. I took a four-pound hammer to the thing. Smashed it to splinters. When the rage passed, I remembered Pirsig’s book. I realized I had never once designed anything I’d built — only ever taken the fastest route to a working result. So I sat down, sketched a proper bed, and built it. A lovely bed that lasted twenty years, and taught me that quality checked for at the end is not quality at all. It has to be designed in from the start.
But that piece never explained how you actually achieve Quality — only that it begins with design and planning before you start building. And that brings us to the subject of this article: Process. Because the secret is that only by viewing everything you do as a process can you formalize it, and only once it’s formalized can you review it at the end of each cycle and ask where the process itself could be improved. This is the entire foundation of what’s called Continuous Process Improvement, or CPI.
I admit, process is an idea which most people find puzzling. For many years, I found myself in difficult situations at cocktail parties when people asked me what I did. I used to obfuscate and say something like, “I’m in software — not lingerie, the stuff in computers.” True, but my real job has always been a process optimizer. Tell that to someone who asks, and they say, “Oh, really? I think my drink needs a refresh.” Hence the ducking and weaving.
But the truth is, I’m a process bigot. Ambrose Bierce defined a bigot as “one who is obstinately and zealously attached to an opinion that you do not entertain.” Mea culpa. I believe everything is a process. And do I hear anyone saying, get real, buddy; what kind of process could this pencil I’m twirling possibly be? For starters, it decays, wears out as you use it and sharpen it or erase marks. Eventually, you throw the stub away, and it decays to dust. Everything is slowly becoming something else, and that is definitely a process.
But to illustrate the true power of process, I’d like to tell you a little story. In the late fifteenth and early sixteenth centuries, the craftsmen who made the swords for the Japanese Samurai had begun experimenting with the type of wood they used to make the charcoal they burned to heat the raw materials for their steel. These swords were fashioned from 200 layers of steel, and thus, to measure the end result, you had to hold every step of the process rigidly the same, without variance, so that you could measure only the difference caused by the change in wood. More than a thousand steps, controlled so tightly that nothing varied. They had, using the process vernacular, an intimate knowledge of their process. That is what tight process control can do for your product or service. Allow you to experimentally vary one thing at a time and measure the result reliably.
And these days, the way to develop this intimate knowledge of any process is to model it. Yes, to literally draw a model of it, and it’s easier than you think because all processes share three things:
1. At least one inflow of something, and you draw it as an arrow and give it a name. A process called bake a cake might label this flow ingredients, or if you want to be more accurate and specific, draw flows for eggs, butter, flour, milk, etc.
2. A process which must be labelled, and which must contain at least one sub-process or ‘recipe for baking a cake’ which explains how you turn the ingredients into the third thing, an…
3. Outflow (or more than one), which might be the finished cake.
That’s it, really, and that’s where my bigotry started. At school, of all places, and specifically on a day in January 1963 when I was fourteen and handed in my South African History assignment.
See, I believe we all have a gift, something we do better than others, and the secret to life is to find yours as early as possible so that you can start to develop it early. And on this day, I discovered that mine was an ability to model things.
My assignment was a half-hull model of Van Riebeck’s ship, The Dromedaris, which he landed in Cape Town harbour in 1652 to establish the first white settlement in Southern Africa. I’d spent three months building this model using balsa wood and glue from a sketch I’d traced at the public library.
The teacher, Mrs. Leaman, took it from me. Shaking with rage, she yelled something at me which I didn’t understand, grabbed my arm and marched me to the headmaster. She’d calmed a little when we reached his office, and she explained, when he invited us in, that I had asked if I could build this model as my assignment and she’d said yes, but look, she continued, thrusting the model at him, it was obviously a store-bought model, and now I was handing it in as my assignment.
I protested to the headmaster that I had made it from scratch. He looked at the model and shook his head, but I said, “I rode to the central library in Johannesburg, found this sketch of the boat’s hull cut longitudinally down the centre, showing all the cabins and decks, but without a rig, and decided it would make a great model. Reproduced it on tracing paper, then transferred the sketch to graph paper and expanded it ten times, then cut balsa wood bulkheads, the frames that formed the side shape of the hull and to which the decks and cabins were attached, and built it piece by piece. I said I had brought the paperwork with me, and we could get it from the class, and he started to smile as he continued turning the ship over in his hands. I added that if you look here, sir, you can see the mistake I made because this part of the deck should be level with this part. He laughed and said, “Wait outside, Mrs. Leaman.”
When she’d closed the door behind her, he said that what I had done was amazing, the best little piece of work he’d ever seen, and that I should be proud of myself. He shook my hand and told me to send “her” in, and when she returned to the class, she was red-faced. Gave me an A minus, and when I looked at her, eyebrow raised, she said, “Well, you did say you’d made a mistake.”
Get Eric Goldman’s stories in your inbox
Join Medium for free to get updates from this writer.
I was cool with that, as it was the first and only time in my entire school career that I earned anything more than a C; it was one of the rare places where my dyslexia didn’t trip me up too badly. Just that one reversal of a measurement. But I realized that day that I could model things very well. Fifteen years later, I started modelling computer systems on a piece of paper using a pencil and eraser, and in 1983 walked into a mall and saw an Apple Lisa and the app called Lisa Draw, which became Mac Draw a year later, and knew I had found a way to replace the paper and pencil. Bought a Lisa ($15,000 Canadian in 1983; my partners almost shot me), and then in 1984, bought a Mac and started taking that with me instead of the pencil and paper, because we now had an early version of the first Deft modeller.
My first startup was a custom software house, and it built the models of systems I designed with our clients; systems that so met their expectations we established a reputation as the best VAX/VMS software house in the world. We literally had companies like Exxon, British Telecom, NASA and Xerox beat a path to our door to ask us to build their systems. All captured in Deft models created in Joint Design sessions with the client, enabling them to see and play with their system before we wrote a line of code.
That is the power of modelling and the power inherent in turning everything into a process which can be studied and improved upon.
Mind you, my trick with the pencil was a bit of a cheat. Physical decay is a process in the sense that entropy is a process — everything is dying, all the time, whether anyone manages it or not. No, that’s a party trick kind of process and nothing more. The process I mean is the deliberate kind — CPI’s maxim is: Think, Plan, Do, Measure and Repeat. That’s a different animal. If you apply it to any process, it will get better and better. Forever.
Robert Pirsig understood all this, years before I ever took a hammer to a bed over it. His book describes two kinds of motorcycle mechanic. One shows up, follows the manual, hits the marks, and clocks out without another thought for the machine — good-natured, competent, and entirely absent; there is, as he puts it, no identification with the job. The other can’t separate himself from the work at all. The bike’s problem is his problem. Its rightness is his own peace of mind. There is, in his words, no other test of the job than the satisfaction it gives you.
Read my bed story again with that in mind. The first bed was built by the spectator mechanic — cut, screw, sand, done, the same shortcut I’d used a dozen times before, because I’d never once stopped to ask if there was a better way. The second bed was the first time in my adult life I was the other kind of mechanic. And that is exactly the piece my original story left out: one afternoon of caring is an event. It is not, by itself, a process. A company — or a life — cannot run on a single well-timed epiphany. The caring has to be turned into something repeatable, or it dies with the mood that produced it.
It’s now time to meet Watts Humphrey, the man who gave this difference a name long before I had the vocabulary for it. Humphrey built what’s called the Capability Maturity Model at Carnegie Mellon — five rungs describing how mature, how genuinely process-driven, an organization’s work really is. At the bottom sits what he called the Initial level: ad hoc, chaotic, success riding entirely on individual heroics, with — and this is the sentence that should worry any manager reading it — no orderly path to improvement even possible. Nothing has been defined well enough to improve. And if you implement this process approach and start maturing, the next level you reach is Repeatable. You can repeat the process and expect the same result, a huge leap above Initial. Then comes Defined, where the process is written down and standardized — no longer one person’s method, but something the whole organization follows consistently. Then you hit Managed, where you start attaching numbers to it: measuring it, tracking its variations, and using those numbers to control it in real time. And once you know what variable changes what, and how, you reach Optimizing: an organization that has built the audit-and-revise loop so deeply into its work that it gets better by habit, not by miracle.
Almost nobody starts above the bottom rung, and very few ever leave it. Disus, by eating its own dogfood, eventually reached the Optimizing stage, producing systems with less than one defect in 100,000 lines of code.
I invented the app called Deft, a computer-modelling tool for designing systems, because the paper and pencil method was too slow and cumbersome. And the tool added rigour and method which improved the models significantly.
My pitch to development shops was always met with, “We’re measured by lines of code, and you want us to spend months on design first?” They were measured on output because output was the only thing anyone could measure. Measuring abstract values is hard, so we often rationalize by treating real, measurable things as if they were all we need. It’s not about you getting what you measure; it’s about learning how to measure what you’d like to get.
The man who saw this most clearly was W. Edwards Deming. In the years after WWII, he tried to sell American industry on almost exactly what I’d later be selling in miniature: that quality has to be designed into a process from the start, tracked, and revised deliberately — and that doing so doesn’t cost money, it saves it. He found real enthusiasm among the engineers who heard him. The managers who ran the plants didn’t believe quality and cost pulled in the same direction.
In 1950, frustrated, he pitched his idea to Japan instead and taught it directly to the country’s top industrialists. When they offered him royalties for the lectures, he told them to fund a prize instead, which is why, to this day, the most prestigious quality award in Japanese industry is the Deming Prize. It would be thirty years before the United States noticed what it had turned away. Only after a 1980 documentary asked, bluntly, “If Japan Can, Why Can’t We?” did American executives go looking for the man they’d shown out of the building a generation earlier.
The heart of what got him laughed at was a simple chain of cause and effect: improve quality, and costs fall, because you’ve stopped paying for rework, waste, and warranty claims. Costs fall, and productivity rises. Productivity rises, and you can compete on quality and price at once. Do that, and you stay in business and keep people employed.
I started this with the fact that I used to tell people I was process optimizer, as I understood that’s what I was doing when I sat with a client and asked what they wanted us to build. We would take their existing processes and automate them; design the way the computer would automate the function, and then look at this potentially automated process and ask, how can we make it more efficient, produce a better quality output, or perhaps produce it in half the time?
And I have now written these two and a half thousand words to try and show you that you can:
1. Treat everything that you do as a process
2. Define the process so that it’s formalized and exists in a place where, at the end of each production cycle, you can compare the result to the process spec, and adjust the spec to eliminate any errors that resulted.
3. Just keep repeating Think, Plan, Do, Measure and Repeat.
If you do all this, you will get better at what you are doing, even if you start as a complete novice.
