The Prototype We Built Twice, And The Notebook I Nearly Threw Away

The skip

The notebook was in a box that was going into a skip.

We were moving out of the unit in Epping, downsizing after a bad year, and I was doing what everybody does in that situation, which is throwing away things at nine at night with a kind of violence that has nothing to do with the things.

It was an A4 hardcover, the ugly green kind, with a coffee ring on the front and the corner chewed by something. It had my handwriting in it from a period of my life I did not particularly want to revisit, and I had it over the skip, and my business partner Dawie took it out of my hand.

He did not say anything clever. He just said, "Not that."

That notebook, eighteen months later, was worth more to us than a piece of equipment we had spent four hundred thousand rand on. I am going to try to explain how, without making it sound like a fairy tale, because it was not one. It was mostly just tedious.

What we were actually doing

We made sorting equipment. Not glamorous. Machines that take a stream of mixed material coming off a conveyor and separate it, and our particular problem, the one that ate seven years of my life, was doing that with a material stream that was wet and inconsistent and that arrived with a lot of contamination in it.

Everybody in that industry knows the standard approach. The standard approach did not work for the specific material we cared about, for reasons that took us two years to properly understand and that I could bore you with for an hour.

So we tried things. That was the whole business, for years. We tried things.

We built a version with an air knife and it worked in the workshop and failed in the field because the field is humid.

We built a version with a different sensor arrangement and it worked brilliantly for eleven days and then the sensor lens fogged and the whole thing was blind.

We built a version around a mechanical solution that I still think was elegant and which shook itself apart in six weeks, and we knew it was going to shake itself apart, we just did not know how fast, and finding out how fast was the entire point of building it.

And then, in year five, we built the thing that worked. And here is the part that matters: the thing that worked was not a flash of genius. It was the fourth idea, informed by exactly why the second and third had failed, and if you had shown me the winning design in year one I would not have believed in it, because I would not yet have known what was wrong with the alternatives.

The word I hated

Nobody in that workshop ever used the word "research".

Research was what happened at universities. Research was people in a clean building with a grant and a laser. We were four guys and a welder in Epping with a shared coffee machine and a leaking roof, and what we were doing was called, in our own mouths, "sorting it out" or "having another go" or, most often, "trying to make the bloody thing work."

I want to be very precise about this because it is the single thing I would go back and tell my younger self.

The language you use in your own workshop determines what you think you are entitled to. We described what we did in the humblest possible words, and then we were genuinely surprised when nobody, including us, ever thought of it as the kind of activity that governments have entire schemes to encourage.

How it came up

Our accountant did not raise it. I want to be fair to her, she was a good accountant, and she did our books and our VAT and she was never once late, and this was simply not her world.

It came up at a braai. It always does.

A guy I had known at university, who had gone into a completely different field, was telling me about a claim his company had made, and I nodded politely for about four minutes and then something in my chest went cold and I said, "Wait. Say that again."

He described what qualified. Not a laboratory. Not a white coat. A genuine attempt to resolve a technical uncertainty where the outcome was not knowable in advance by a competent professional in that field.

And I sat there with a beer in my hand and I thought about a machine that shook itself apart in six weeks.

The part where I nearly ruined it

I went in the next week and I told Dawie, and we got excited, and then we did the thing that I now know is the single most common mistake, and which I have since watched two other companies make.

We went back and tried to write the story afterwards.

We sat in a room for two days trying to reconstruct, from memory and from invoices, what we had done in years two and three. And it was hopeless. Not because we could not remember roughly what we had tried, but because we could not evidence it. We could not say when a decision was taken, or what the alternative had been, or what the specific uncertainty was that we had been trying to resolve, or why the result had surprised us.

An invoice for steel tells you that you bought steel. It does not tell you that you bought steel because a hypothesis about vibration damping had just been disproved.

So we sat there with a pile of invoices and a very sinking feeling, and Dawie got up and went out to the storeroom and came back with the green notebook.

What was in it

Everything.

Not because I had been clever. I had kept that notebook because I am a slightly anxious person who does not trust his own memory, and because my first boss, an old toolmaker, had told me at nineteen to write down what I did each day and why, and I had done it out of obedience for a year and out of habit for the next fifteen.

It had dates. It had the arguments. It had entries like "Tried 4mm at 30 degrees, no better than 6mm, so it isn't the angle" which is not a triumphant sentence, it is a man closing a door, and it turns out that a record of a door being closed is exactly what this whole business runs on.

It had a page where I had written, in capitals, "WHY IS IT ONLY WET ONES", underlined three times, which I do not even remember writing, and which turned out to be the beginning of the thing that eventually worked.

When we finally sat down with the R&D Tax consultants that a friend put us onto, the woman went through that notebook for most of an afternoon with a pencil, and at one point she stopped and looked up and said, "You do understand what you have here."

I said, honestly, no, not really.

What I learned about the process, which is not what I expected

I had assumed the whole thing would be about money. Receipts, timesheets, cost allocation.

It is not. The money part is the easy part and any competent bookkeeper can do it.

The hard part, the part that everything turns on, is the technical narrative. Somebody has to be able to explain, in writing, to a person who is intelligent but is not an engineer in your field, what the uncertainty was, why a competent professional could not simply have looked up the answer, what you did about it, and what you found out.

And here is the cruel part. The engineer who did the work is usually the worst person in the building to write that document. I know, because I tried. I wrote eleven pages and they were unreadable. I assumed context. I skipped over the interesting bits because they were obvious to me. I described the machine that worked and barely mentioned the three that did not, because I was, still, at forty two years old, embarrassed about them.

The failures are the claim. That took me a long time to accept. If the thing works first time, you did not resolve an uncertainty, you just did your job. The six weeks where the machine shook itself apart is the most valuable six weeks in the entire history of the company, and I had spent years describing it as the year we wasted.

Australia

I ended up in Australia in 2019, which is a whole other story involving a divorce and a job offer and a decision I made in an airport that I do not fully understand to this day.

I work for a mid sized manufacturer in Adelaide now, in engineering, and about a year in I sat in a meeting where a young design lead was describing a program of work that was, unmistakably, exactly what we had done in Epping. Trial, failure, redesign, failure, redesign.

I asked, afterwards, in the corridor, whether they claimed for it.

He said he thought that was a thing for tech companies.

I said the same sentence to him that had been said to me at a braai, and I watched his face do exactly what my face did. And when their finance director eventually brought in an R&D Tax specialist, the first thing that person asked for was not the accounts. It was the engineering logs. Which did not exist, because nobody had told those engineers that their notebooks were worth anything.

They started keeping them the next week. That is the only piece of consulting I have ever done in my life and I did it in a corridor for free.

Dawie, and Manchester

Dawie is in the north of England now. He married a woman from Manchester and he does something in industrial automation that he has explained to me four times.

He phoned me last year, at a completely unreasonable hour, because he never checks the time difference, and told me that his current company had been running an eighteen month development program on a control system and had never claimed a cent, and that he had, in his words, "gone full Epping on them."

He told me he had stood in a meeting and told a room of English engineers about a green notebook and a coffee ring, and that they had looked at him like he was mad, and then their finance people had gone and looked at what the R&D Tax Incentive rules there actually say, and had come back rather quiet.

Three countries. Same conversation. Same look on the same kind of face.

What I actually think

I am suspicious of stories where somebody discovers a scheme and gets rich. That did not happen to us. What we recovered was real and it was not life changing, and it went, in the least romantic way possible, into paying down a facility and buying a second hand lathe.

What changed was smaller than money and bigger than money.

It was that I stopped being ashamed of the failures.

For seven years I had carried the private conviction that a better engineer would have got there faster, and that the three machines that did not work were evidence against me. And then a woman with a pencil sat in our office for an afternoon, and turned the pages of my ugly green notebook, and treated every single failure as a valuable, deliberate, professional act.

Nobody had ever done that. Not my partner, not my accountant, not my father, and certainly not me.

The notebook is in a drawer in Adelaide. I have not written in it for years, because I have a new one, and the new one is exactly the same, and I write in it every day.

Date. What I tried. What I expected. What actually happened. Why I think that is.

It takes me four minutes. My old toolmaker boss died in 2011 and I never told him what that habit was worth, and I think about that more than I should.