Cases — Innovation Atlas

#	Case	When	Mech	For	Con	Obs	Gen	Yield	Link
1	The 1956 AT&T consent decree and compulsory licensing draft	1949–1961	M06	5	5	4.5	5	22.5	free
2	Lean's Engine Reporter and open benchmarking in Cornwall draft	1811–1904	M05	5	4	4.5	5	21.0	free
3	Samuel Insull and the invention of the utility business model	1892–1932	M01	5	5	4.5	4	21.0	free
4	The Electric Home and Farm Authority: financing the demand	1933–1942	M03	5	3	5.0	4	20.0	free
5	Oral rehydration and the door-to-door campaign draft	1968–1990	M03	5	5	4.0	5	20.0	free
6	The German dye industry invents the research laboratory	1863 – c. 1900	M01	5	5	4.0	5	20.0	free
7	The heavy plough and the shape of northern Europe draft	c. 900 – 1300	M04	4	5	4.5	4	19.5	free
8	The Great Fire and the invention of fire insurance draft	1666 – 1774	M10	4	4	4.5	5	19.5	source
9	The Plimsoll line draft	1870–1890s	M10	4	4	4.5	5	19.5	source
10	The air brake and the law that finally deployed it draft	1869–1900	M02	5	4	4.0	5	18.7	free
11	Abolishing local time draft	1840 – 1884	M03	4	5	4.0	5	18.7	source
12	The pallet, the forklift, and the war that standardised them draft	1930s–1950s	M01	3	4	5.0	4	18.3	free
13	The Alcorn County cooperative draft	1934–1940	M07	4	3	5.0	4	18.3	free
14	Phosphorus, and the ore that Bessemer could not use draft	1856–1885	M01	4	4	4.5	4	18.0	source
15	The Cistercians as a technology transfer network draft	1098 – c. 1300	M03	4	3	4.5	5	18.0	source
16	The wind-powered sawmill and the fluyt draft	1594 – c. 1670	M05	4	4	4.5	4	18.0	source
17	The Amsterdam Wisselbank draft	1609 – 1795	M01	4	4	4.5	4	18.0	source
18	Song dynasty iron and the industrialisation that did not follow draft	c. 1000–1250	M02	4	4	4.5	4	18.0	source
19	The Casa da Índia and navigation as a state knowledge system draft	1500–1600	M03	4	4	4.5	4	18.0	free
20	Penicillin: the half of the story that produced the drug draft	1928–1945	M01	5	5	3.5	5	17.5	source
21	Pacioli did not invent double-entry bookkeeping	c. 1300 – 1494 – 1600s	M07	5	5	3.5	5	17.5	free
22	The barcode deadlock and how it broke draft	1970–1985	M03	4	4	4.0	5	17.3	source
23	Mittasch's catalyst campaign at BASF draft	1909–1913	M01	4	5	4.0	4	17.3	source
24	The Venetian Arsenal draft	c. 1104 – 1600s	M03	4	4	4.0	5	17.3	free
25	Denmark's Heat Supply Act and planned district heating draft	1979–present	M10	4	4	4.5	3	16.5	source
26	The Baťa system at Zlín draft	1900–1939	M01	4	3	4.5	4	16.5	free
27	Japan's Top Runner standards draft	1998–present	M01	4	4	4.5	3	16.5	free
28	Danish wind turbines against the American great leap	1975–1995	M05	5	4	3.5	5	16.3	source
29	The Xerox 914 and charging by the copy draft	1959–1970s	M01	5	4	3.5	5	16.3	free
30	Negative learning by doing in the French nuclear programme draft	1974–2000	M07	5	5	3.5	4	16.3	source
31	The rail gauge that the Romans did not set	c. AD 100 – 1846	M07	5	4	3.5	5	16.3	free
32	Wichterle's hydrogel lenses and the value that left the country	1953–1971	M06	4	4	4.0	4	16.0	source
33	Los Molinos: two years of persuasion, almost no adoption draft	1950s	M02	5	2	4.0	5	16.0	free
34	The stirrup thesis draft	8th century, argued 1962	M07	4	3	4.0	5	16.0	source
35	The gauge war draft	1838 – 1892	M03	4	4	3.5	5	15.2	free
36	Midgley: two successful innovations, two catastrophes draft	1921–1974	M09	5	5	3.0	5	15.0	free
37	Movable type existed for four centuries before Gutenberg draft	c. 1040 – 1450s	M03	5	5	3.0	5	15.0	free
38	Domesday counted the mills draft	1086	M01	3	3	4.5	4	15.0	free
39	The Chappe optical telegraph draft	1794 – 1855	M02	3	3	4.5	4	15.0	free
40	Kalundborg: industrial symbiosis nobody designed draft	1961–present	M05	4	3	4.0	4	14.7	source
41	Power by the Hour draft	1962–present	M01	4	3	4.0	4	14.7	free
42	Interchangeable parts took fifty years, not one demonstration draft	1798 – c. 1855	M07	5	4	3.0	5	14.0	source
43	Deming, quality, and the method that emigrated draft	1950–1980	M06	4	5	3.0	5	14.0	free
44	The Lunar Society draft	1765–1813	M03	3	3	3.5	5	12.8	free
45	Containerisation and the rebuilding of everything around a steel box	1956–1980	M03	5	5	2.5	5	12.5	source
46	OGAS: the Soviet network that was not built draft	1959–1970	M02	3	2	4.5	3	12.0	source
47	The Kaplan turbine and low-head hydro draft	1912–1930s	M01	3	4	4.0	2	12.0	free
48	Roman concrete and the thousand-year gap draft	c. 200 BC – AD 1750s	M02	4	4	3.0	4	12.0	source
49	How solar became cheap draft	1957–2020	M04	5	5	2.5	4	11.7	free
50	Nokia: distributed attention and shared fear draft	2005–2011	M07	4	4	1.5	4	6.0	source
51	Kodak, correctly told draft	1975–2012	M07	4	4	1.0	3	3.7	source

CASE-0005 · M01 Invention is not innovation · 1892–1932

Samuel Insull and the invention of the utility business model

Chicago, United States · bears on SRC-0002, SRC-0014 · supports

21yield · force 5 · obscurity 4.5

What happened: Insull arrived at Chicago Edison in 1892, aged 32, after being passed over for the presidency of Edison General Electric. He concluded that the cost of electricity was governed less by fuel than by load factor — the ratio of average to peak demand — and therefore by who the customers were and how they were charged. On a visit to Brighton in 1894 he found Arthur Wright's demand meter, which recorded not only how much a customer used but how fast. Insull licensed it and introduced two-part rates in Chicago: a fixed charge for the capacity held open, and a very low charge per unit consumed. He then deliberately assembled customers whose peaks fell at different hours — streetcars in the morning, factories through the day, homes at dusk — and pursued industrial users who were self-generating. In 1898, to the National Electric Light Association, he argued that exclusive franchises should be traded for state rate regulation, at a time when most of his peers opposed it. In 1903 he committed Chicago Edison to a 5,000 kW Parsons steam turbine at Fisk Street, far larger than anything then in service, against a reluctant board.
What the popular story gets wrong: There is barely a popular story to get wrong, which is the point. Edison and Tesla are household names; the man who designed the industry's commercial and regulatory structure is not. Where Insull is remembered at all it is for the 1932 collapse of his holding-company empire and the fraud trial that followed — he was acquitted after two hours of jury deliberation — rather than for the forty years of structural design that preceded it.
What it bears on: Supports SRC-0002 (Drucker) on industry-structure change as a source of opportunity — Insull created the structure rather than responding to it. Supports SRC-0014 (Breznitz) on the stages after novelty being where value is captured: he invented almost no technology and captured an industry.
Implication for us: Almost every structural feature of the business we work in was designed by one operator between 1892 and 1912, which means those features are choices and not physics. When we treat tariff design, the obligation to serve, or the demand charge as fixed background, we are treating a hundred-year-old commercial decision as a law of nature. The two-part tariff in particular was invented to solve exactly the problem flexibility products face now.

CASE-0008 · M03 System, not gadget · 1933–1942

The Electric Home and Farm Authority: financing the demand

Tennessee Valley, United States · bears on SRC-0001 · supports

20yield · force 5 · obscurity 5.0

What happened: By 1935 roughly 90% of American urban homes had electricity and roughly 10% of rural ones did. Having secured the mandate to generate and distribute cheap power, TVA director David Lilienthal identified the binding constraint as neither generation nor lines but the household balance sheet: a refrigerator cost a substantial fraction of annual family income, utilities would not cut rates until demand grew, and demand could not grow until appliances were affordable. He called it a complete business stalemate. The Electric Home and Farm Authority was created by executive order in December 1933, managed initially by the TVA directors, to finance appliance purchase and home wiring — small deposit, terms up to five years, a low annual charge — while negotiating with manufacturers for cheaper models sold through local utilities and cooperatives. In 1935 it moved to the Reconstruction Finance Corporation and expanded beyond the Tennessee Valley alongside the Rural Electrification Administration.
What the popular story gets wrong: The New Deal electrification story is told as dams and transmission lines. The dams are photogenic and were not the constraint. The instrument that actually converted infrastructure into consumption was a consumer credit agency, and it is almost entirely absent from the popular account.
What it bears on: Supports SRC-0001 (Rogers) on relative advantage as something the customer must be able to realise, not merely perceive: the advantage of electricity was obvious and unreachable without the appliance. Qualifies any account of adoption that treats price per unit as the decision variable.
Implication for us: The direct modern parallel is heat pumps, home storage and EV charging, where the constraint is capital cost to the household and not generation, tariff or device performance. Any adoption forecast we produce that models unit price and not household financing is modelling the wrong variable. The historical answer was an instrument, not a campaign.

CASE-0040 · M01 Invention is not innovation · 1863 – c. 1900

The German dye industry invents the research laboratory

Germany · bears on SRC-0015, SRC-0022 · supports

20yield · force 5 · obscurity 4.0

What happened: Perkin's mauveine (1856) was British and so was much of the early chemistry, but the industry was built in Germany. Bayer was founded in 1863, BASF in 1865, alongside Hoechst, Agfa and Cassella. Through the 1870s these firms did something new: they built permanent in-house research laboratories staffed with doctorate-holding chemists. Hoechst set up an analytical laboratory in 1870 and a research organisation by about 1880; BASF's laboratory took shape from 1874 under Heinrich Caro and had five or six academic chemists by 1887, with a central research laboratory designed later that decade; Bayer built the Elberfeld scientific laboratory under Carl Duisberg. The 1877 German patent law sharpened the incentive by making chemical inventions defensible. The scale of commitment is visible in BASF's synthetic indigo programme: research from 1880, commercial production in 1897, seventeen years and roughly two million pounds. By 1913 Germany was exporting on the order of 135,000 tonnes of dyes against Britain's 5,000.
What the popular story gets wrong: Two things. First, the story is usually told as chemistry — Perkin's accident, then a march of molecules — when the decisive move was organisational. Second, the origin of the industrial research laboratory is popularly assigned to Edison at Menlo Park or to Bell Labs. Historians of the field have located it in the German dyestuffs firms, and Bell Labs is downstream of a decision taken in a dye works.
What it bears on: Supports SRC-0015 (Knott): R&D is an organisational capability with measurable returns, and the returns here were captured by the firms that built the capability rather than by the country that made the discovery. Supports SRC-0022 — this is exactly the sort of finding that exists in the history-of-technology literature and is absent from the trade-book canon.
Implication for us: The lab we run is an instance of a form invented in the 1870s for a specific purpose: to convert academic training into defensible patent positions in a science-based industry. When we ask what our lab is for, the honest first answer is to check whether that original purpose still describes us, and to say so if it does not.

CASE-0030 · M07 The famous story is wrong · c. 1300 – 1494 – 1600s

Pacioli did not invent double-entry bookkeeping

Italian city-states, then Europe · bears on SRC-0005, SRC-0012 · qualifies

18yield · force 5 · obscurity 3.5

What happened: Luca Pacioli's Summa de Arithmetica (Venice, 1494) contains the first printed description of double-entry bookkeeping — 27 pages inside a 615-page mathematics textbook. He described a method Italian merchants had already used for roughly two centuries; he did not invent it. What the Summa did was couple an existing practice to a new distribution technology: printing. Werner Sombart argued in 1916 that capitalism was unthinkable without double entry. Basil Yamey spent decades arguing the opposite from the ledgers themselves, showing merchants rarely used it to compute periodic profit at all. The dispute is unresolved and worth presenting as unresolved.
What the popular story gets wrong: Pacioli is routinely called the inventor of double-entry bookkeeping and the father of accounting. He was neither. He described, in print, a method Italian merchants had been using for roughly two centuries — de Roover traced the practice through medieval merchant account books well before 1494. Pacioli's own contemporaries were not confused about this; the attribution hardened later.
What it bears on: Qualifies SRC-0005 (Arthur): the recombination here was of a practice with a printing press, not of two technologies. Supports SRC-0012 (Godin) obliquely — the misattribution of an anonymous collective practice to a single named author is precisely the kind of tidying that conceptual history exists to undo.
Implication for us: Write down the unglamorous thing. The most consequential document in the history of business administration was a chapter inside a maths textbook, describing what practitioners already did, and it worked because it was printed. Codification of existing practice is a legitimate and undervalued form of innovation — which is, incidentally, what this repository is.

CASE-0002 · M05 Bricolage beat breakthrough · 1975–1995

Danish wind turbines against the American great leap

Denmark, United States · bears on SRC-0003, SRC-0005, SRC-0018 · supports

16yield · force 5 · obscurity 3.5

What happened: Two national wind programmes ran in parallel after the oil shocks. The United States funded large, aerodynamically ambitious machines designed largely by aerospace engineers and built to specification. Denmark's turbines came out of agricultural machinery workshops and small firms, were smaller and cruder, were sold to local co-operatives who operated them, and were revised continuously against field failures. The Danish design lineage converged on the three-bladed upwind configuration that became the world standard; the American programme produced machines that broke and a domestic industry that did not survive its own subsidies.
What the popular story gets wrong: It is usually filed as a subsidy story or a national-character story. The mechanism was neither. It was the feedback loop: distributed co-operative ownership put operators, failures and designers within one short cycle of each other, so learning accumulated in the design lineage rather than in reports.
What it bears on: Supports SRC-0003 (von Hippel) — the Danish machines were substantially user-shaped, with operators inside the design loop. Supports SRC-0018 (Edmondson) on intelligent failure: the Danish programme's failures were small, survivable and acted upon, which is the only configuration in which failure is worth anything. Qualifies any reading of SRC-0005 that treats recombination as a purely cognitive act — here it was organisational.
Implication for us: When choosing between a specified-then-built pilot and an instrumented-then-revised pilot, the second wins whenever the failure cost is survivable. Our gate criteria should therefore ask for the revision cycle time, not only the design.

CASE-0029 · M07 The famous story is wrong · c. AD 100 – 1846

The rail gauge that the Romans did not set

Roman Britain, north-east England · bears on SRC-0001, SRC-0005 · contradicts

16yield · force 5 · obscurity 3.5

What happened: The story runs: Roman chariot ruts fixed cart axle widths, wagonways copied them, Stephenson copied the wagonways, and 4 ft 8½ in is therefore a Roman measurement. It is false, and it has been debunked since at least 1905. North-east English wagonways used no standard at all — Wylam and Blenkinsop's Middleton line ran at 5 ft, Beamish at 4 ft 4 in, Bigges Main at 5 ft 7½ in. Stephenson used 4 ft 8 in because that was Killingworth colliery's gauge where he was engine-wright, added a half inch around 1829, and the 1846 Gauge Act made it law only because eight times more track already existed at his gauge than at Brunel's 7 ft. The real convergence is duller and more interesting: Roman rutways and Northumbrian wagonways both had to fit a cart drawn by two horses in harness, roughly 1.4 m. Same constraint, no inheritance.
What the popular story gets wrong: Everything causal in it. The claim that Roman chariot ruts fixed cart widths, which fixed wagonway gauges, which fixed Stephenson's gauge, has no supporting chain. British wagonways used no common gauge: Wylam and Blenkinsop's Middleton line ran at 5 ft, a Beamish wagonway at 4 ft 4 in, Bigges Main at 5 ft 7½ in. Roman traffic was overwhelmingly pedestrian and pack, and war chariots had been militarily obsolete long before the Roman period. The Railway Magazine was already debunking the tale in 1905, which tells you how durable a good story is. The NASA solid-rocket-booster coda that usually closes the telling is a later embellishment on an already false premise.
What it bears on: Contradicts nothing in Bedrock directly, and that is its value: it is a case about how we reason. It qualifies SRC-0005 (Arthur) usefully — recombination and inheritance are real mechanisms, and this is a case where a plausible inheritance story turns out to be independent convergence on a shared physical constraint.
Implication for us: Two rules. First, path dependence is a real mechanism and therefore an attractive explanation, which makes it exactly the kind of claim to check rather than repeat. Second, when two systems share a dimension, look for a shared constraint before assuming one copied the other — the answer here is horses, not Romans.

CASE-0003 · M06 Value captured elsewhere · 1953–1971

Wichterle's hydrogel lenses and the value that left the country

Czechoslovakia, United States · bears on SRC-0014, SRC-0015 · supports

16yield · force 4 · obscurity 4.0

What happened: Otto Wichterle developed hydrophilic gel polymers suitable for implantation and, after being removed from his institute post, produced the first usable soft contact lenses at home over Christmas 1961 on a spinning apparatus assembled partly from a child's Merkur construction set. The rights were subsequently sold abroad by Czechoslovak state organs on terms that returned very little to the country or to Wichterle. The global soft-lens industry was built elsewhere on the result.
What the popular story gets wrong: It is told domestically as a story of ingenuity, and it is one. The consequential part is the second half: a country can hold world-leading invention and capture almost none of the value, and the mechanism of that loss is institutional rather than technical.
What it bears on: Supports SRC-0014 (Breznitz) directly and painfully: novelty-stage invention without the stages that follow it produces prestige and no industry. Provides a domestic case for SRC-0015 (Knott) on the difference between generating knowledge and appropriating returns from it.
Implication for us: Every concept in the lab should carry an explicit appropriation question at the concept gate: if this works, who captures the margin, through what mechanism, and what would have to be true for that to be us. A concept that cannot answer it is a gift to somebody.

CASE-0001 · M03 System, not gadget · 1956–1980

Containerisation and the rebuilding of everything around a steel box

US, then global · bears on SRC-0001, SRC-0005 · supports

12yield · force 5 · obscurity 2.5

What happened: Malcom McLean, a trucking operator with no shipping background, ran the first container voyage in 1956. The metal box itself was trivial and not new. What took the following two decades was everything around it: purpose-built ships, gantry cranes, redesigned ports, container-compatible chassis and rail cars, the ISO dimensional standard, renegotiated dock labour agreements, and rewritten insurance and customs treatment. Cargo handling costs collapsed, and the geography of manufacturing reorganised itself around the result.
What the popular story gets wrong: It is told as an invention story with a clever outsider at the centre. The box was the cheapest component. The innovation was a two-decade renegotiation of a socio-technical system in which most of the necessary work was standards-setting, capital reallocation and labour politics — none of which appears in the telling.
What it bears on: Supports SRC-0001 (Rogers) on compatibility as an adoption constraint: containers were radically incompatible with the installed base, and the diffusion curve reflects the cost of removing that incompatibility rather than any property of the box. Also supports SRC-0005 (Arthur) on recombination — every component existed.
Implication for us: Our own candidates are mostly of this shape: a plug-in battery, a flexibility product, a metering interface. The device is the cheap part. Before a gate, state explicitly which standards, tariffs, certifications and counterparties would have to move, and treat that list — not the hardware — as the project.

Ranked by what you don't already know.

The index

Written up