What Is Invention?

This is a co-authored post with Jarmo Markula

"Invention" is one of the most confidently used words in the business of growth, and one of the least examined. It appears in strategy documents, policy frameworks and pitch decks as though everyone agrees what it means. They do not. The word has been quietly doing the work of two other things — innovation and discovery — for long enough that the distinction has been almost entirely lost. That loss has consequences, now legible in slowing productivity, persistent gaps in the pipeline, and the particular kind of frustrated ambition that comes from trying to commercialise things that were never properly invented.

Joseph Schumpeter was clear on this in 1911, and clarity has not improved with the intervening century. Invention, he wrote, is the act of conceiving something new: the first demonstration that an idea is possible. Innovation is the act of arranging the economic requirements to implement it, the first commercial exploitation of an invention (aka the realisation of new value). Diffusion is what happens after both. Three stages. Three distinct acts, with different actors, different logics, and different timelines. He was not describing synonyms. He was describing a pipeline.

The distinction matters because the stages are not interchangeable. What it takes to invent something is not the same as what it takes to commercialise it. An inventor, in Schumpeter's sense, is someone who can work productively with a raw, unproven idea before there is any legible commercial signal — someone comfortable, as my co-author Jarmo Markula recently put it, with outliers, unconventional people, and the kind of knowledge that has not yet been made useful. An innovator, by contrast, needs a legible signal. Their job is to arrange resources, build markets, and move fast. The two roles require different skills, different risk tolerances, and different institutional environments. Giving the same word to both has made it easy to assume the same people and institutions can do both.

For a specific window of time, in a specific institutional form, something close to the real thing existed. Bell Labs at its peak in the late 1960s employed fifteen thousand people, fourteen of whom went on to win Nobel Prizes. Its researchers gave the world the transistor, the laser, the photovoltaic cell, Unix, optical fibre, and the cellular network. Xerox PARC, which carried the torch into the 1970s, added the graphical user interface, Ethernet, and the laser printer. A telephone company and a photocopier maker between them defined the information age — not because they were doing research in the academic sense, and not because they were scaling companies in the startup sense, but because they were doing the work that sits between those two things. Patient, multi-disciplinary, commercially insulated, tolerant of raw ideas and unconventional approaches. They were, in Schumpeter's sense, inventing.

We have retained the word and removed the conditions. And so we go on trying to commercialise things that, in Schumpeter's sense, have not yet been invented.

That institutional form has largely gone. Arora, Belenzon, Patacconi and Suh, writing in Innovation Policy and the Economy (University of Chicago Press, 2020), documented the structural shift: the past three decades have been marked by a growing division of labour between universities, which focus on research, and large corporations, which focus on development. Corporate publications per firm fell by around forty per cent between 1980 and 2015. Bell Labs was separated from AT&T in 1996. Xerox PARC was spun off as an independent company in 2002. DuPont's Central Research laboratory — in operation since 1903 — was closed in 2016. The same paper noted that a scientific idea now requires "additional integration and transformation to become economically useful," and that small firms and university technology transfer offices cannot fully substitute for what the great corporate labs had provided: multi-disciplinary work at the scale required to solve significant technical problems.

What filled the gap? Universities do research, under a publish-or-perish incentive that rewards novelty and priority over development and depth. Startups do commercialisation, under a time-to-market pressure that rewards speed over the kind of patient iteration that turns a raw idea into a useful one. Neither operates in the middle. The result is the gap that policy has spent decades trying to close: the valley of death, the space between what a laboratory produces and what an entrepreneur can actually build with. Research into the valley of death has repeatedly found that even where governments have substantially increased financing and management support for this transition phase, the challenge of crossing it persists. Adding money to a people problem does not make it a resource problem. It makes it an expensive people problem.

The people who are missing are what Jarmo and I describe in this piece as invention actors: those who can take what science produces and do the integration and transformation that Arora and his colleagues describe. Not researchers. Not entrepreneurs. The people in between, who need institutional protection from commercial pressure precisely because the value they are generating is not yet legible enough to be funded on commercial terms. Bell Labs produced them because it was structured to produce them — with long time horizons, multi-disciplinary teams, and the insulation that came from being embedded in a regulated near-monopoly with patient capital. No current institution is structured that way, and the modern innovation system has no category for what they do.

This is the covert work the word is doing. "Invention" appears throughout the policy and strategy literature as though it describes something that is happening, when what it mostly describes is something that happened, in a specific institutional context that has since been dismantled. We have retained the word and removed the conditions. And so we go on trying to commercialise things that were never properly invented — not because the underlying science is absent, but because the stage that would turn science into something economically ready has no institutional home.

Restoring it will not be a matter of more grants, more accelerators, or more technology transfer offices. It will require rebuilding the category: organisations structured for the specific work of invention, sitting between basic research and commercial development, staffed by people who can hold unconventional ideas long enough to make them useful. The transistor took from 1947 to the late 1950s to become the thing everything else runs on. That is not a pipeline failure. That is what invention actually looks like, done properly, in institutions that understood what they were doing.

The word has been waiting, with some patience, for the institutions to catch up with it.

Special thanks to Jarmo for his help in initiating the article, we may well be back with more!

Until next time, Editor of Ideas for a Better World

Jarmo Markula is a Strategic Reinvention Partner to Corporate, City & Ecosystem Leaders. Helping leaders build structures that unlock continuous growth. The concept of "invention actors" and the observation that the Schumpeterian pipeline has been collapsed rather than merely slowed I credit to him, from our discussions.