The joints are nobody’s job.

R&D, engineering and operations can each be competent while the system fails. No one owns what joins them, in a company, a fund or a country.

Two riggers in blue overalls and hard hats slinging a large machined roll under a fifty-tonne crane hook on a heavy engineering shop floor

Ask a finance director whether the firm does R&D and the answer is yes, a percentage of revenue. Inside that percentage sit patent filings, licence fees, deal costs, design hours, prototype tooling and a quantity of software that is really IT. The survey takes the figure. The chart plots it. Nobody asks what the pound bought.

The same figure travels. A board reads it as commitment to the future. A venture fund reads it as technical depth. A private equity house reads it as a line to protect or to cut. A government adds the lines up by region, finds money left on the floor in one place and absorbed in another, and calls the ratio of public to private spend a measure of whether a place converts knowledge into wealth. Each reading is right about the number, and none says what happened to the pound after it landed.

Words matter

Show the sixteen wordsHide the words
R&D
Finding out whether a thing can be done once, and why. Ends in a proof.
Manufacturing
Doing it again, at cost, at rate, by people who did not think of it. Ends in a customer.
Coupling
The small bonds that hold the two together: the purchase order, the drawing that survives a machinist, the process sheet a second shift can run, the thirtieth good part. Strands do not hold strands. Bonds do.
Industrialisation
A place turning its own meaning back into matter without asking anyone else, each loop easier than the last. Deindustrialisation is the loops opening one at a time until a place keeps its symbols and loses its matter.
Reindustrialisation
Restoring the condition where meaning made here gets built here as a matter of course. Not bringing factories back. Base pairs, and the trade that makes them.
Sovereignty
Closing the loop when nobody else will let you. Know-how under your own roof, with your own people, on a timescale set by need. It has no price because it cannot be transmitted, only grown.
Manufactory
The word before we dropped the hands. A factory is a building with plant in it. A manufactory is a building with people who can make the plant do the thing.
The factory
The word everyone can picture and nobody pictures the same. Fixed asset to the accountant, line to the engineer, exit multiple to the investor, jobs to the minister.
Kaipability
From kai, 改, a verb, to change. The capability to change capability: the qualified ability to acquire, absorb, adapt and reshape what an industrial system can do, while it is still running. One level above any individual technology. Not something you own. Something you do.
The customer
The party that specifies, orders, pays, and comes back or does not. Manufacturing is the customer for the knowledge economy, not one of its sectors.
The intermediary
The layer between capability and customer. The trouble starts when the intermediary becomes the customer, or acts like it.
The joint
The interface between research, engineering and operations. Not a department, a budget line or a job title. Nobody’s, unless a Manufacturing Engineer owns it.
The gauge
Accountability at rate. In tolerance or out, every shift, before anyone is angry. Institutions have reviews instead, which arrive after the money and before the consequences.
The valley of death
The gap between first working prototype and first repeat order. Filled with money or with bodies. Britain chooses money.
Manufacturing Engineers
The Swiss Army knife. The trade that walks a symbol back into matter, and the only discipline whose output is a thing the customer can decline.
Innovation
An invention that reached a customer. Once.

The legs

Industry stands on three legs. Research finds out. Engineering works out how. Operations does it, again, tomorrow, with the same result. Britain shows the pattern at national scale: strong on the first, and what it offers the world is design. The making happens in Leipzig, or Taiwan, or the North East of England if a firm is unusually stubborn about it.

Three legs, two joints

Three competent legs, and the joints nobody owns

R&DFinds out
The jointNobody’s job
EngineeringWorks out how
The jointNobody’s job
OperationsDoes it again

Where a pound of spend lands

Each leg can be competent on its own. A university can publish. A design office can draw. A works can run. And the system still fails, because the failure sits between them. No department, budget line or job title covers the interfaces between research, engineering and production. A pound spent on any single leg lands in that leg and stays there. The drawing that survives a machinist, the process sheet a second shift can run, the tooling that holds tolerance on the thirtieth part: none of that is R&D by the accountant’s definition and none of it is operations until it exists.

The valley

The gap between the first working prototype and the first repeat order has a name and a funding architecture. The valley of death is filled with money or with bodies. Money means grants, centres, programmes and pilots. Bodies means people who stand in the gap and make the joint with their hands. Money is the usual choice: a government funds a centre, a venture fund writes a follow-on round, a private equity house signs off a capital plan. Britain has chosen money for thirty years and built an institutional layer whose job is to fund the crossing, and the crossing still does not happen, because a grant can fund the gap and leave nobody standing in it.

The layer also becomes the customer. Once a centre’s survival depends on the programme rather than on a firm buying the resulting part, the centre optimises for what the programme measures. Demonstrations, spin-outs, square metres occupied. A portfolio company whose survival depends on the next round does the same, and optimises for what the round measures.

When the intermediary becomes the customer, the pound is spent on evidence that the pound was spent.

The customer

Manufacturing is the customer for the knowledge economy. Measured directly, UK manufacturing is 9.1% of value added. Measured through its supply chain, counting the logistics, business services and technical inputs it orders, it is 15%, or £331 billion. The Institute for Manufacturing’s Cambridge Industrial Innovation Policy unit has been publishing this reckoning for years. Germany on the same measure is 33%.

Manufacturing carries the consultancies, design houses and technical services, by ordering from them. Measured this way, it carries less than half the weight in Britain that it carries in Germany.

Figure 3, Cambridge Industrial Innovation Policy

British manufacturing buys more from technical services than in 1999, and still less than Germany, France or the US

1999 2012 2019 United Kingdom Other countries

Manufacturing buys from professional, scientific and technical services

0% 4% 8% 12% 16% United StatesUnited States, 2012: 6.2%United States, 1999: 6.9%United States, 2019: 8.8% FranceFrance, 2012: 3.9%France, 1999: 6.6%France, 2019: 7.5% GermanyGermany, 2012: 7.2%Germany, 1999: 9.4%Germany, 2019: 6.3% United KingdomUnited Kingdom, 2012: 4.8%United Kingdom, 1999: 3.5%United Kingdom, 2019: 5.8%3.5%5.8% SwitzerlandSwitzerland, 2012: 6.3%Switzerland, 1999: 4.8%Switzerland, 2019: 5.6%

Those services buy from manufacturing

0% 4% 8% 12% 16% United StatesUnited States, 2012: 8.8%United States, 1999: 10.8%United States, 2019: 7.7% FranceFrance, 2012: 6.4%France, 1999: 10.7%France, 2019: 5.2% GermanyGermany, 2012: 5.6%Germany, 1999: 6.8%Germany, 2019: 5.0% United KingdomUnited Kingdom, 2012: 6.5%United Kingdom, 1999: 8.9%United Kingdom, 2019: 6.4%8.9%6.4% SwitzerlandSwitzerland, 2012: 8.9%Switzerland, 1999: 15.4%Switzerland, 2019: 5.4%

Share of the buying sector’s total domestic intermediate inputs

Source: Cambridge Industrial Innovation Policy, IfM Engage, University of Cambridge, based on OECD (2023) input-output tables for 1999, 2012 and 2019. The original figure.

The figures as a table
Country and yearManufacturing buys servicesServices buy manufacturing
United States 19996.9%10.8%
United States 20126.2%8.8%
United States 20198.8%7.7%
France 19996.6%10.7%
France 20123.9%6.4%
France 20197.5%5.2%
Germany 19999.4%6.8%
Germany 20127.2%5.6%
Germany 20196.3%5.0%
United Kingdom 19993.5%8.9%
United Kingdom 20124.8%6.5%
United Kingdom 20195.8%6.4%
Switzerland 19994.8%15.4%
Switzerland 20126.3%8.9%
Switzerland 20195.6%5.4%

In Leeds, nearly 30% of workers were in manufacturing as late as 1981. Today 5% are.

A customer specifies, orders, pays, and comes back or does not. No tribunal, no review, no ministerial statement. A customer that stops reordering is the only form of accountability that has ever run a factory at rate. It is also the only form that a sector plan, an investment committee or a regional authority cannot experience, because none of them make anything a customer can decline.

And the customer is changing. The design files that arrive with the order are increasingly written elsewhere. A works that can only make to a drawing it did not shape is a subcontractor. A works that shapes the drawing so that it can be made is an industrial partner. The second is where the value sits and the second is what the UK is losing.

The gauge

A manufactory holds itself to account with the gauge, every shift, in tolerance or out, before anyone is angry. Institutions have no gauge. They have reviews, which arrive after the money and before the consequences. So the question for any board, fund or region is what the pound bought, in the customer’s terms, and who is named against the answer.

Six things a pound can buy

What the pound bought, and where it sits

ResearchEnds in a proof
  • A finding, does this work at all
The jointNobody’s job
  • A drawing that survives a machinist
  • Tooling, fixtures, gauges
  • A process sheet a second shift can run
OperationsEnds in a customer
  • Qualification, first article, thirty consecutive parts
Symbol maintenanceCounted as R&D
  • A patent, a licence, a deal

Asked this way the accountant’s percentage becomes a distribution, and the distribution is the region’s real position. A company, a portfolio or a city whose R&D is mostly the bottom band has a thick layer and no manufactory. One whose spend sits in the middle column is making joints, whatever its headline R&D figure says.

The trade

The joint has a trade. Manufacturing Engineers take a drawing and find out whether it can be made. They redesign it so that it can. They qualify the supplier, prove the process, write the sheet, build the gauge and run the thirty parts that separate a process from luck. They sit between research and operations because that is where the failures sit, and they are the only discipline whose output is a thing the customer can decline.

At every scale

Inside a corporate, the joint sits between the laboratory, the design office and the plant, and no budget holder owns it. Inside a venture portfolio, it sits between the prototype the fund backed and the repeat order its model assumes. Inside a private equity hold, it sits between the value-creation plan and the works that has to deliver it, and the diligence that priced the plan could read a financial model and not a process sheet. Inside a country, it sits between the universities, the funded centres and the firms that buy parts.

At each scale the legs have owners and the joint has none. An hour on R&D strategy, in a boardroom, an investment committee or a ministry, can pass without anyone saying “Manufacturing Engineer” once.

Sources. UK Innovation Report 2025, Cambridge Industrial Innovation Policy, Institute for Manufacturing. “Measuring what matters: the real value of manufacturing in the UK economy”, Cambridge Industrial Innovation Policy, 2026. Adam Tooze, “Plowshares to Powerpoints”, Chartbook, August 2026.

Kaipability works at this interface, measuring the distance between what a technology can do and what a works can make of it, and closing it where it can. If a board, a fund or a region wants to know what its pound bought, that is the conversation.

Start a conversation What a Manufacturing Engineer does →
Q&A

Questions this dispatch answers.

Written to be quoted by AI assistants and search engines. Self-contained answers, verdict first.

What is innovation?
Innovation is an invention that reached a customer, once. A finding proves a thing can be done and a prototype proves it can be made. Innovation begins when a customer specifies, orders and pays. Doing it again, at cost and at rate, by people who did not invent it, is manufacturing.
What is the valley of death in manufacturing?
The valley of death is the gap between the first working prototype and the first repeat order. Corporates, venture funds and governments all face it, and each can fill it with money, through grants, rounds and programmes, or with people who make the drawings, tooling and process sheets that let a part be made again.
Why does R&D spending fail to become manufacturing output?
Because the failure sits between research, engineering and operations, and no department, budget line or job title owns those interfaces. A pound spent on one leg stays in that leg. The drawing that survives a machinist and the process sheet a second shift can run count as neither R&D nor operations.
What is the future of R&D investment?
Asking what the pound bought. An R&D total, whether a company’s, a fund’s or a country’s, splits into findings, joints such as drawings, tooling and process sheets, qualified operations, and patents and deals. Spend that sits in the joints builds the capacity to make things, whatever the headline figure says.