The meeting room had been in session for days.
The subject was localising equipment.
The boss picked a valve up off the table, held it in front of the foreign executive sitting opposite, and asked one question:
"A valve exactly like this. Why can't we use it?"
The foreign executive gave the reason on the spot.
The valves on the boat are bronze — a high corrosion class, they will not rot in seawater, and once fitted they last for years. The one in his hand is brass, a domestic part, bought from a hardware shop to connect a water pipe. The two look almost identical. Material, standard and corrosion performance are entirely different.
The boss turned the valve half a turn in his hand and put it back on the table.
He said nothing more.
This starts before the factory was built.
Before 2008 the boss went to the United States and looked at a great many yachts, moored in row after row, and what he was looking at was how big the market was.
When he came back the decision came quickly: a joint venture with Robertson & Caine of South Africa, fifty-fifty, build the factory in China.
As to where the things on the boat would come from, his answer was simple.
You can buy anything in China.
No need to check, no need to ask. China has the fullest range of industries and can make anything. The things on a boat are valves, pipes, lights, appliances — aren't they?
That conclusion was reached before the feasibility study. It did not even occur to him to ask: what if we cannot buy something?
He looked at the market half. He did not look at the supply chain half — because he assumed there was nothing to look at.
And one more thing that nobody at the time thought was a problem: in the whole team, nobody had ever built a boat.
Not management, not the shop floor, and not our purchasing department.
The factory was built in 2008, and localisation started in the first week.
The boss had come out of kayaks. Composites were his own trade. What went into a hull, and whether it could be sourced in China, he knew.
Resin first. They had been using Scott Bader from the UK; we changed to DSM. The unsaturated resin for the hull, the vinyl ester for the outer layers, all changed. Fibreglass cloth and core material followed, one by one, into Chinese equivalents.
Changing a material still needs a signature. The factory had an engineer over from the United States who knew composites — what could be substituted and what could not, he could decide on the spot.
One who knew the trade, one who could sign. That is how the material layer moved forward.
With that layer, the hulls could be built.
Hull after hull came out of the workshop.
And then they stopped.
Some days later someone finally asked: what about the things that go on the boat?
Where the engines were, where to buy them, when they would arrive — nobody had raised any of it before.
It was not that nobody could answer. The engineer sent by the foreign side had never raised it. The people who came over from South Africa had never raised it. The workshop, purchasing and management had never raised it.
Not one of those items had been bought.
A single boat has over five hundred items of assembly material, involving eighty-four suppliers.
We finished building the hulls before we asked the question for the first time.
Isn't that strange?
The judgement that "you can buy anything in China" was a little over a year old.
Reality does not argue. It simply put something nobody had predicted onto the table.
Localisation half succeeded and half failed.
The dividing line is very clear: whether the item existed in China.
The composite layer did. Resin, fibreglass cloth, balsa core, gelcoat — all available in China, all workable. The boss had come out of that trade, the American engineer could sign off, and the actual work was our purchasing department running it down item by item. When a problem was thrown up, someone could always catch it.
Further up: engines, navigation, sanitary ware, deck hardware. Those did not exist in China. Nobody understood the design, and nobody could say "this substitute will do".
Somebody has to sign off on whether a substitute goes on the boat. And whoever signs carries the responsibility.
But to sign, there has to be something sitting there first. If there is not even an item in the country, nobody can sign off on anything.
At the boat level, nobody could catch it. The boss's background did not help here, and neither did ours.
That valve in the meeting room is where it came from.
In the spring of 2009 the boss insisted on sending people over to see.
Four went: me, an interpreter, an electrical man and an assembly man.
Nobody in China understood this, so we had to go and learn.
Robertson & Caine had been building boats in Cape Town for twenty years. Of the world's catamarans built each year, more than thirty percent come out of Cape Town, and more than eighty percent of that from this one company. Only there did anyone know where the things on a boat come from.
The bill of materials they sent was entirely in English, with a few established abbreviations. No standard part numbers, no drawings, no photographs — just a bare table.
With a table like that, nobody in China knew what to buy.
Later I worked out why. Purchasing development in South Africa was done by the technical department. Technical agreed the specification with the supplier, generated an abbreviated code in the system, and purchasing ordered against the code.
That works in South Africa, because those suppliers had been partners for ten or twenty years.
In China, an abbreviated code on its own is nowhere near enough.
I visited close to forty suppliers, going boat by boat against the physical goods to find where each material was fitted and how. Every week I wrote a report and sent it back.
In May 2009 I submitted a summary report.
It set out three options: best, middle, worst.
The best was this: do not build our own system; order through theirs. Orders placed in the South African system, with the material data, prices and suppliers inside it, and North American and European suppliers shipping straight to China.
But that had a precondition: South Africa had to agree.
The middle option: no access to their system; order from our own spreadsheets. Five hundred-plus items all done by hand, and several times the headcount.
The worst option: build a domestic supplier system of our own.
Why did I put that one last?
Because it was not realistic.
First, the industrial base was not here. At the 2008 Dutch marine equipment show there were well over twelve hundred exhibitors; only twelve were Chinese, and they were casting stainless steel hardware — a very small part of a boat's materials. More than half the exhibitors were concentrated in the United States, Britain, France and Italy.
Second, certification was out of reach. Things on a boat do not simply go on once they are made. Critical valves, cable, engines, life-saving equipment and navigation equipment all sit in the classification society's certified catalogue: either factory approval or type approval, with drawings to be reviewed and type tests to be run, and the certificate expires. Nobody in China was doing that whole set at the time.
Third, China was more expensive. For the same item, domestic channels quoted eighty percent above South Africa. One Velcro nylon fastener: 8.04 yuan a metre in China before tax, 4.47 in South Africa. One marine antifouling paint: the Hong Kong general agent was seventy-five percent above South African local. A Plastimo liferaft from France: 5,927 US dollars through the domestic agent, 1,954 direct — a spread of more than double.
Localisation carries a default assumption: that domestic is cheaper. What we found was the opposite.
When the report went in, the company's answer was to go in two stages.
Stage one, about three years, mainly overseas purchasing, with domestic suppliers developed in parallel. Stage two would be mainly domestic.
At roughly the same time, at the show in South Africa, the foreign side's purchasing director made it plain: purchasing development work had to be carried out under South African direction throughout, and had to stay in step with South Africa.
One side told us to build our own system. The other told us not to.
Both sentences were addressed to us.
I have thought about this many times since.
A complete feasibility study is normally broken into a few blocks: market, technical, financial, organisational — and supply chain.
The first blocks answer "is this worth doing at all". The supply chain block answers a different class of question. The standard framework in the industry breaks it into five:
Is the source of supply available?
Can lead time and delivery be guaranteed?
Is there a single source?
Make, or buy?
How are logistics and customs to be arranged?
These five are the standard subjects of a feasibility study.
Now put those five back onto this project and go through them one by one.
All five. Not one had an answer.
The market half asks whether anyone will buy. The supply chain half asks whether anyone will sell to you.
Only when both halves are done is it a feasibility study.
I have remembered that valve ever since.
It looked exactly like the one fitted on the boat.
What was missing was not the grade of the copper. It was the chain that takes decades to grow — material, process, standard, certification, each link inside the next, all in someone else's hands.
So the question was never really about a valve.
It was: what is this boat going to be filled with, and on what.
That question should have been answered before the joint venture agreement was signed.