
An INTP man’s answer to a badly framed question is a better question, and organizations do not buy questions. Fifteen years of him being right about the framing, and fifteen years of a file explaining that he cannot hold a date, are the same fifteen years.
In 2013 a customer wanted to know whether we could turn their torque wrenches around in three days instead of five, and I put the question to Vaughn Steckel in front of six people, and what I got back was a question about what a day was.
“Three days from when?”
I said from receipt.
“Receipt at the dock or receipt at the bench? Because there’s a day in between and it isn’t ours.”
I said we’d count from the dock, like everybody else.
“Then we can’t do three and neither can anybody else, and the ones telling them three are counting from the bench.” He looked at the ceiling for a second. “We could do three from the bench and tell them four from the dock, which is honest and sounds worse.”
Delphia Mott quoted them four. They took it. They are still a customer.
That is the good version of the thing this piece is about, and I have told it as a good story, and I want it understood that I have spent fifteen years living on the other side of that same faculty and calling it a problem, and being right about it.
Vaughn came to Hesse in 2004. He was twenty-eight, physics degree, three years at an instrument maker in another state where his job had apparently been to sit in a room and work out why gauges disagreed with each other. He is a heavy, quiet man who reads while he eats and has never once brought a lunch that required a fork. He does not decorate. His bench is a bench. He will answer a yes-or-no question with a paragraph and he knows that he does it and has apologized for it perhaps three times in twenty-two years, always in the middle of doing it again. In twenty-two years no one else I have employed has come to me unprompted to say that a thing he did five years earlier was probably wrong.
Cass Doheny came to me in his third year with a repeatability set that was not repeating and asked whether the gauge was bad.
I sent him to Vaughn, which is what everybody does.
He came back an hour later with the answer, which was that the gauge was fine and that Cass had been approaching the setpoint from the same direction every trial, and that a set taken that way tells you about the operator.
I asked what Vaughn had said.
“He asked me what I thought repeatability was,” Cass said. “Then he waited.”
That is what it is like to work near him and about half the people who have worked here have loved it.
Hesse Metrology & Calibration is in Redfern in a building that was a bakery. Twenty-two people. We do dimensional, torque, pressure, temperature, and a little electrical, mostly for valve and fitting manufacturers and two aerospace suppliers who keep us honest. Rutger Hesse started it in 1981 out of a garage with a set of gauge blocks and a customer list of nine. I came in 1994 as a technician and bought it from him in 2007.
There is a sentence I have said about Vaughn Steckel to a bank, to two customers, and once to a room of my own supervisors. I made it. I stand behind about half of it.
He’s the smartest person in this building and you can’t hand him anything with a date on it.
Our accreditation reassessment is in three weeks. In the pre-visit review the assessor flagged that our uncertainty methodology — the document everything we sell rests on — lists an author who does not appear on our authorized signatory list. That is Vaughn. He has never been a signatory, because signatories commit to turnaround, and I have fifteen years of records explaining why he does not commit to turnaround, and I wrote most of them.
Rutger, who is eighty-eight and comes in at the holidays, has a line he still uses about the lab and it has always been about Vaughn without ever saying so.
“The certificate’s a piece of paper. What are you selling?”
The Budget
A calibration certificate is a single page carrying a number, and the number is not the point.
Anybody with a decent standard can compare a customer’s gauge against it and write down what the gauge read. What makes that comparison mean anything is the uncertainty — a statement of how well we know the thing we are telling you. And nobody looks the uncertainty up. You build it. You sit down and enumerate every contribution to the doubt: the standard’s own uncertainty, its drift since it was last certified, the resolution of the instrument, the repeatability across a set of trials, the temperature of the room, the technician’s alignment, the fit of the mathematical model to the data. Each of those gets a value and a distribution, and they get combined, and the combination gets expanded to a level of confidence, and what comes out is a number that goes on the certificate in a line most customers never read.
Two labs can measure the same gauge, write down the same reading, and one of those certificates is worth something and the other one is worth nothing, and there is no way to tell them apart by looking.
Which means the entire product is a claim about what you do not know. That is the trade. And the man who is best in this building at enumerating what we do not know is the man this lab has never put on a schedule.
Nobody opens the north door. There is a sign, and the sign is not the reason. The dimensional room holds twenty degrees and the north door opens onto the loading side, and one door held open for a minute costs you the better part of an hour of stabilization on the CMM. Cass Doheny propped it in his second week to bring in a pallet and Delphia said one sentence to him that he has never repeated to me and he has been here nine years and he has never done it again.
The dimensional room also has a log. Temperature and humidity, every hour, on paper, initialled, going back to 2009, in a binder that has never once been looked at by anybody except an assessor and Vaughn, who reads it the way other men read a paper.
I am a licensed auctioneer. I do two charity auctions a year and I am very good at it, which surprises people who have met me at work, and my wife’s position is that it is the only two days a year I am recognisably enjoying myself.
Rev 0
The document the assessor flagged is fourteen pages and it is called the uncertainty estimation procedure and it was issued in 2009 and its revision history has one line in it.
Rev 0. Seventeen years.
We were going for accreditation in 2008 and I had a consultant and the consultant had a template, and the template was a template — it produced a document that would pass and that did not describe what this lab actually did. Vaughn asked whether he could have a go at it. I gave him what I thought was a generous window and he came back inside it, and I have gone and checked the dates because I did not believe my own memory: eleven weeks.
He has never revised it. Not because he is precious about it. Because nothing has come up that it does not already handle. We added pressure in 2012 and temperature in 2016 and the procedure absorbed both without an amendment, because it was not written about torque, it was written about what a measurement is, and the extra scopes were instances.
Two assessors in a row have said, in the closing meeting, that they wished more labs had one like it.
And everything else he has ever touched has grown. A two-week job becomes six weeks and the six weeks are all necessary and the customer does not care. I can produce the list. The temperature scope took nine months because the interlaboratory comparison he insisted on turned up a bias in a reference bath that we then had to chase to the manufacturer, and he was right, and we could have had the scope in nine weeks with a bath nobody had checked.
Ida Osterhout ran dimensional here through the back half of the last decade and had the identical instrument and used it the same way. When Vaughn opened a job up it was called rigour and it was said with a certain amount of pride, the way you talk about a thoroughbred that will not go in a trailer. When Ida did it she was told twice, by me, that she needed to be more pragmatic, and once that she was overthinking it, and she is a technical fellow at a much larger organization now and I read something she wrote last year that I did not fully understand. What the identical habit costs a woman in a lab like this one, and how much faster it costs it, is set down on its own page.
Twenty-two people. Three of them women. I have hired eighteen of the twenty-two.
Ida came to see me before she went and told me she’d had an offer.
I asked what it would take.
“Nothing you’d want to give me,” she said, and then she said she meant that literally and not as a dig, and I have never settled whether it was true.
Well, There You Go
The thing that made him easy to beat, and I did beat him, is that he supplies the counter-argument himself.
He will raise an objection — a real one, the kind that saves you nine months later — and then, because the counter to it has also occurred to him and because he does not distinguish between arguments that help him and arguments that do not, he will say the counter out loud, immediately, in the same breath. Although if the drift is linear over the interval that mostly goes away. There is no concession in it. He is finishing the thought. He has, in his own head, put the objection and its answer on the table and is waiting to see which one survives.
A room does not hear that. A room hears a man withdraw his own point.
In 2013 — this is the one — we were deciding whether to bring the pressure scope in-house or keep subcontracting it, and Vaughn objected on the basis that our proposed reference would be at the wrong end of its range for half the work we actually get, and then he said, in the same breath, that you could probably cover the low end with a second standard if you wanted to spend the money.
And I said, “Well, there you go.”
And went on to the next item.
There was a laugh. It was not a cruel laugh and I did not intend a cruel thing; I intended to keep a meeting moving, and it moved. We bought the one standard. Four years later we bought the second one, at a higher price, after two customers complained about the low end.
He said nothing to me about it then or ever. What he did do, and it took me about six years to notice, is stop putting the counter-argument out loud in meetings. He still had it. He would come to my office afterward and give me both halves, alone, where the second half could not be used on him.
I had trained him to withhold the honest part of his thinking from the room and I experienced that as him becoming more effective.
Delphia noticed before I did. She said something to me in about 2019, in the kitchen, that I did not take up.
“He used to argue with himself out here.”
I said he still does, in my office.
“That’s what I mean,” she said.
Last Look
By about 2015 the arrangement had settled into a shape that nobody proposed.
Vaughn did not get jobs. Vaughn got last look. Somebody else would take a piece of work — a new scope, a customer’s technical complaint, a method change — and drive it to a decision on a date, and then, before it went out, we would take it to Vaughn.
He would find something. He found something every single time, for a decade and more. Sometimes it was small and sometimes it stopped the thing dead, and it was always real.
The last look on the aerospace scope in 2017 took forty minutes and I have a note of it because I made one.
He read the method through once without saying anything, which takes long enough that people fidget, and then asked whether the resolution figure we had used was the display resolution or the least significant digit we could actually resolve.
Delphia said they were the same thing.
“They’re the same number,” he said. “They’re not the same thing. If the last digit’s noisy you can’t claim it.”
We went and looked. It was noisy. The scope went out a week later with an uncertainty about a third larger than the one we had been about to publish, and that is the version an aerospace supplier looked at and accepted, and the smaller one would have been challenged inside a year by somebody who checks.
It is a magnificent use of a man and it is also the reason his name is not on anything. A last look is not authorship. You cannot put it on a signatory list, you cannot bill it, and at the end of a decade of it the honest summary of his contribution in any document I would show a bank is that he reviews other people’s work.
In 2018 I did not take him something.
We were re-establishing the calibration interval on a family of pressure transducers for a customer who wanted to go from six months to a year, and it was a straightforward reliability argument, and it was late, and I knew — I knew — that Vaughn would want to look at how the historical data had been collected before he would agree it supported anything. Delphia and I did it without him and I signed it.
The interval went to a year. Eight months later that customer had a transducer come back at three times tolerance with no way to bound when it had gone, and the reason was that half the historical data we had leaned on had been taken as-found and half as-left, and we had pooled them.
Which is exactly the thing he would have asked about, in the first five minutes, in the form of a question about what a day is.
I told the customer it was a process failure and I told my own people it was a process failure, and it was a process failure, and I have never once said in any room that the process failure was me routing around a man on purpose because I did not want a week gone.
Vaughn read the corrective action when it went round. He came to my office and said the root cause statement was fine as far as it went.
I asked how far that was.
“To the pooling,” he said.
Ottoway
Our biggest account is Ottoway Flow Control, and in 2018 one of their master torque standards came back to us reading four per cent low.
An out-of-tolerance finding on a master goes backwards. Everything that master was used to calibrate since its last good certificate is now suspect, and everything those tools were used to assemble is now suspect, and Ottoway assembles valves that go into places where a bolted joint matters. Their quality manager, Ilya Braun, called me at seven in the morning and by nine there were nine of us in the front room of a bakery in Redfern with a whiteboard.
We spent an hour and ten minutes and got nowhere, and The reason, stated flatly: nobody in that room could state the question. People were arguing about the recall window and about whether the drift was linear and about who was going to call the customer’s customer, and every one of those arguments was downstream of something nobody had settled.
Vaughn was not in the room. He had been in the lab with the master and the data since six.
He came in at about ten past ten with two pages and did not sit down.
“You’re asking the wrong thing,” he said. “You’re asking what’s affected. Ask what tolerance the affected work actually needed.”
Ilya said all of it needed to be in tolerance.
“No, the tools needed to be in tolerance. The joints needed a clamp load.” He put the first page on the table. “Four per cent low on the master means the wrenches were set high. Everything assembled with them is over-torqued, not under. For your product, under is the failure mode. Over is a fatigue question and it’s a different question and it’s smaller, and for most of these joints it’s inside the design margin.”
The room went quiet.
He put the second page down. It was a list of the specific part families where over-torque was not inside the margin. There were three of them.
Ottoway recalled work from three lines instead of everything, and it took nine days instead of the quarter everybody in that room had been assuming, and Ilya Braun has told that story at two industry meetings, and I have heard him tell it, and he tells it as a story about his own team.
Ilya Braun asked me on the phone that afternoon who the fellow with the two pages was.
I said that was Vaughn, and that he had been here since 2004.
“Since 2004,” Ilya said. “We’ve been a customer since 2006.”
I said he mostly works in the back.
I have gone back over it a good deal since — that the man whose entire reputation in this building is for making things bigger walked into the worst morning of my career and made it smaller in ninety seconds.
The Training File
Under our standard we have to keep records demonstrating the competence and training of personnel. Every person here has a training file and the files get pulled at every assessment.
In 2016 our quality manager at the time was building the training matrix and the matrix looked thin, and she put everyone through an online course called team effectiveness that took two hours and cost nine dollars a head, and the completion certificate from that course includes a results page.
Four letters and a paragraph. Filed, with a document number, under a retention schedule.
Vaughn’s file has his degree, his instrument-maker employment verification, our internal competence sign-offs by scope, and that. He has never taken another course, in twenty-two years, because he has never encountered one that taught him something he did not already have.
His page says INTP.
I read the whole thing this spring, including the appendix, which is where the course put the part with any content in it. It says the letters compress an order of operation: the habit doing the heavy work, the habit it goes out into the world by, and, last, two that a man reaches for on purpose or not at all, slowly, and badly under pressure. Which habit sits where for this one is written up on the page the type gets to itself, out of the sixteen that course was working from.
The habit under the load is a model of how a class of thing works, built privately, maintained privately, and answerable to nothing except whether it holds together. He does not ask whether the number on the certificate is right. He asks what would have to be true for that number to mean anything, and then goes and finds out whether those things are true. The model was never about a torque wrench. It was about what a measurement is, which is why the procedure he wrote in 2009 absorbed two new scopes without an amendment.
There is a neighbouring version of this man who runs the same habit on the object in front of him and stops when the object behaves. That one is a diagnostician and he is finished when the machine runs. This one runs it on the reasoning, and reasoning has no surface you can put your hand on, so there is nothing that tells him he is done.
The second habit is why. It generates alternatives, continuously, unbidden — including alternatives to whatever he has just said. That is the well, there you go mechanism and it is not a weakness of character, it is the same faculty that produced the second page in the Ottoway meeting, which is a list of exceptions to his own reassurance. In a room, honesty of that kind is indistinguishable from retreat.
And the two at the back of his order are the two a date is made of. Knowing what is good enough for the person who has to use the thing — which is a judgment about people and not about the thing. And carrying the concrete accumulated record of how long work of this sort has actually taken here, in a form you can reach for. A commitment to a date is a claim made in the absence of complete knowledge, on the strength of precedent and of a read on what the other party will accept, and he has very little of either, and he has a great deal of the faculty that keeps finding reasons the claim is premature.
So the sentence I built about him is accurate. Read a second way, it is a description of a man missing exactly the two things that convert knowledge into a delivery, and this lab decided that the missing two were the whole person and used the rest of him at the end of other people’s processes for a decade.
The necessary part, since I am the one who ran it this way. If you are asked to release a result you do not stand behind, or to narrow a notification a customer is entitled to, that is not a personality conflict, and two people at a whiteboard are not the forum for it — accreditation bodies have complaint channels, customers have contractual rights, and in most places a person who raises it has protection. And the other end of it, which is mine. Fifteen years of being told you cannot work to a deadline, with nobody ever writing down what on time means for your role or what training was offered, is a description and not a performance conversation, and you are entitled to ask for the second one in writing.
The course’s own materials say the questionnaire is for self-awareness and has no business in staffing decisions, that about a third of people sitting it a second time a month later land somewhere else on the grid, while the scales it is cut from move much less.
At our 2019 assessment the assessor asked what the team effectiveness record demonstrated.
The quality manager said it demonstrated training.
He wrote it down and moved on.
The Signatory List
I have to answer the assessor’s finding in three weeks and there are two ways to do it.
I can restate the uncertainty methodology under a current signatory’s name, which is a paperwork exercise and which is what my consultant has recommended, and which would take an afternoon.
Or I can add Vaughn to the signatory list, which requires a documented competence justification, and a documented commitment to turnaround, and a supervisor’s statement, and the supervisor is me.
Two things came loose after Ottoway and I have not got either back where it was.
The torque wrenches in 2013, first. I have told that one for thirteen years as a charming story about a man who cannot answer a simple question. What he did in that room was find the ambiguity that every competitor in this market was exploiting, hand me an honest number, and then hand me the alternative — three from the bench, four from the dock — which is a fully worked commercial option, delivered in about nine seconds, by a man I was already describing as unable to commit to anything.
Then the pressure meeting, and well, there you go. I have carried that as a small piece of meeting management. It was the moment I taught a man that being honest about the weakness in his own position would be used against him inside a second, and he learned it, and after that the second half of his thinking came to my office in private, where it was worth a fraction of what it was worth in the room.
The things the file has no room for: he has been in a small claims dispute with a neighbour over a fence since 2021 and has represented himself and is, by his own account, losing on grounds he considers incorrect; he has a sister in Tacoma he calls every week and has never mentioned to anybody here; and of everybody who has worked here he is the one who once asked me how the business was actually doing, meaning the balance sheet, and he asked it in 2011 and sat through the whole answer.
The competence justification is a form. There is a box on it for the supervisor to describe the basis for the determination.
I opened the file this morning. What I put in the box was that the candidate authored the laboratory’s uncertainty estimation procedure, which has required no revision in seventeen years.
That is true and it is one sentence and it is not a commitment to turnaround, and the form has another box under it, and I have not filled that one in.
The people in these essays are composites drawn from long observation. The theory is not.