question
the next disruption is not a better technology. it is a different value system.
A new technology typically reaches people by one of two routes, or a blend of them: incumbents absorb it, or newcomers displace them. Its existence delivers no value. Use does, and early use delivers only part of it.
Where products depend on shared standards or on other users, an unresolved contest between the routes can delay broad use for years.
Our proposition is that it can be shortened. One actor, the initiator, changes its own payoff logic and builds an inclusive ecosystem on shared infrastructure, the flywheel, in which joining becomes progressively cheaper than resisting. Incumbents and newcomers tend to meet the same six shifts on the way, and each actor that makes them lowers the cost for the next.
That is what this manifesto is about. It does not pick a route. It asks how the value reaches people sooner, whichever route prevails.
Innovation labs, hackathons and transformation roadmaps cannot change the outcome if they leave the underlying incentives untouched.
For builders, investors and institutions prepared to go first.
ecosystem-value-first. initiator-led. open-by-design.
how.this.manifesto.is.built
Four parts, each answering one question. Each answer raises the next.
What this builds on. Most of the mechanisms here are established.
- Rogers described how innovations spread through a population over time.
- March showed why organisations favour exploiting what they know over exploring what they do not.
- Christensen showed why incumbents’ own incentives hold them back.
- Cohen and Levinthal showed that firms can only use outside knowledge they are equipped to absorb.
- Edmondson showed that people raise problems only where it is safe to.
- Arkes and Blumer showed that people keep investing in what they have already paid for.
- O’Reilly and Tushman described why exploring and exploiting need separate structures.
- Chesbrough described how useful knowledge crosses company boundaries.
- Ostrom described how shared resources can be governed without a single owner.
- Farrell and Klemperer set out how switching costs and network effects lock markets in.
- Stigler set out how regulators come to serve the regulated.
- Wasserman set out how rarely founders stay in control.
What is ours. Four things are our own proposal and still need testing: that the recurring blockers group into six shifts; that incumbents and newcomers largely share them; that a transition can start with one actor changing its own payoff logic; and that a four-stage flywheel can make joining cheaper than resisting. We state them as propositions.
Two words carry the argument. An actor is anyone in the system: a person, a company, a fund, a regulator, a community. The initiator is the actor that changes what it optimises for and moves first.
four.premises
Premise 1. There are two archetypal routes.
In adoption, incumbents absorb a technology into the existing market structure. In disruption, newcomers displace that structure.
From here on the word means this second route only; the headline uses it in its everyday sense. Most real transitions are hybrids, in which incumbents absorb some of the technology, newcomers take part of the market, and in some fields both advance side by side. The contest is over which route dominates, and on whose terms. This manifesto takes no side in it.
Premise 2. Existence delivers no value. Use does.
The internet, the touchscreen and the neural network each existed for years before their value was unlocked.
Invention creates an option and use exercises it. Early users capture part of the value. Most of the rest arrives when use becomes broad.
In markets with strong network effects or incompatible standards, unresolved competition can delay that.
Premise 3. A long contest is costly for most of those involved.
While the routes compete, buyers wait, suppliers hedge, and capital and effort go into the side that loses.
Not every contest is wasted: a short one can reveal the better design, and settling too early can lock in a worse one. The waste begins when continued competition teaches less than it costs in duplication, incompatibility and delay. These costs are highest where products depend on shared standards or on other users, the markets economists describe in terms of network effects and switching costs. Where several solutions can coexist, the contest is cheap and this manifesto has little to add.
Premise 4. The same things hold both routes back.
Once a technology works, most of what delays its use sits outside whichever actor carries it: adopters’ switching costs, missing complements, standards nobody will adopt alone. These bind incumbents and newcomers alike.
Inside the organisation the biases are human, and newcomers that grow take on incumbents’ inertia.
What follows from the four.
A contest of this kind tends to end in one of three ways. One side wins outright, often with a system it controls. A regulator mandates an answer, which is quick but can fix the wrong one. Or the parties converge on infrastructure that none of them controls.
Our central proposition is that an actor can bring about the third ending deliberately, and that it is the faster one whenever joining costs a participant less than continuing to resist.
six.shifts
Every organisation has a value system: a working set of beliefs about what is worth doing, what gets rewarded and what counts as success. Most are well tuned to defend the existing business, because that is what they were built to do.
The six shifts are our model of where such value systems most often need to change. Each has four parts: a default setting inside an actor, the blockers it tends to produce, a design rule that counters it, and a first move any actor can make. The first four act inside organisations and the last two across markets. They are recurring patterns, not a checklist: a given transition may hinge on two of them and barely touch the rest.
The actors behind the blockers are not villains. They respond sensibly to the incentives in front of them, which is why culture programmes alone rarely hold. Incentives, measures, governance and the flow of knowledge have to change with them.
Status follows seniority, certainty and a record of success.
Opinion as religion.
When the most senior person’s view settles a question, testing stops. Internal consensus drifts away from what is happening outside, and because disagreeing is costly, the drift goes unreported.
Default to “prove me wrong”. Before any significant decision about a market, a technology, a direction or a hire, ask what would have to be true for it to be wrong. If nothing could, it is a belief. If the answer is specific and testable, it is the next experiment. Make the question a governance requirement, and give credit to people who change their minds on evidence.
Before agreeing with the most senior voice in the room, ask what would make it wrong. If you cannot answer, say so. For an organisation: stop letting rank settle open questions.
People are rewarded for hitting targets, and a failed project damages a career.
Sunk cost inertia.
Past investment creates pressure to continue whatever the evidence says, and the pressure grows with the size of the investment. The sunk cost effect is a well-documented cognitive bias. Leaving it unchecked is a governance failure: set the stop criteria before the money is spent, and let someone who did not make the original call decide whether to continue.
A team that framed a clear hypothesis, tested it properly, got a definite no, stopped and shared what it learned has done valuable work. Tata’s Dare to Try award and NASA’s Lean Forward; Fail Smart award exist to recognise that.
Pay for evidence. Reward honest early failure. A deliverable should be able to fail a test: could data that might realistically exist prove it wrong? A strategy deck or market analysis that could not is an opinion, and should be priced as one. Pay for prototypes, experiments and tested hypotheses, including those that came out negative. Reward the quality of the thesis and the cleanliness of the exit, not failure as such.
Publish one failure with its evidence: the thesis, what you tested, what broke.
We trust our customers and our current market data, and optimise for both.
Metric misalignment. Innovation theatre.
A new model tends to look worse on the existing metrics before it looks better on any new one. Quarterly revenue, headcount efficiency and satisfaction scores for existing products were designed for the current business, so they register the new mainly as a cost. Labs, hackathons and innovation titles can absorb the pressure to change without changing anything, because they are measured on inputs such as ideas submitted and workshops held.
Give the new its own scoreboard. Measure the new by what it proves, keep that scoreboard apart from the core metrics, and count outputs. Give non-customers and edge cases the same attention as the customers already served, because that is usually where the new first shows up.
Report one signal the current system does not measure.
Whoever built it keeps running it, and governance exists to keep things consistent.
Founder bottleneck.
The skills and roles a company needs can change sharply as it scales. Where no handover happens, a role can outlive its phase. The same holds for a seed-stage board kept through later rounds, or an early colleague kept in a role out of loyalty.
In Wasserman’s study of 212 American start-ups, fewer than a quarter of founders were still chief executive at the IPO. That shows how common the handover is; it does not show the cause.
Set the direction, rotate the roles. Give a clear Mission and wide latitude on how to reach it. Let roles change hands as the phase changes, on triggers agreed in advance, and put governance on rotating terms.
Put the Mission ahead of your current position, and name it clearly enough that handing over a role does not feel like abandoning it.
Knowledge is a competitive advantage, to be held and protected.
Ecosystem closure. Technology lock-in. Regulatory capture.
Platforms often open up to attract users and builders, then restrict access and raise prices once they dominate. Once enough connections, data and workflows are built on a system, the cost of switching can exceed the gain from a better alternative. Regulators can come to protect the firms they oversee, and compliance costs that incumbents can absorb become entry barriers for everyone else.
Build neutral ground early, and keep it contestable. Openness is not the goal in itself. A tightly governed system can sometimes coordinate better than an open one, and a nominally open one can still be captured. What matters is that shared infrastructure stays contestable: others can interoperate with it, take their data with them, replace any component including the operator, and leave at a tolerable cost, under governance that no single actor controls. That is what open means in this manifesto. Set it up before large commercial interests form. The period in which a technology is proven enough to attract serious investment but not yet entrenched is short, and separating infrastructure from its owner afterwards is possible but far more expensive.
Share what you have learned before anyone asks.
Uncertainty is a risk to be removed before acting.
Nash equilibrium lock. Academia-practice gap.
Each actor in a market optimises its own side of the transaction. The result can be a stable state that nobody has a reason to leave alone, even though most would gain if all moved together. Academic incentives concentrate rewards around publication, and industry incentives around proprietary application. Neither side is consistently rewarded for turning controlled evidence into independent, reusable proof in practice, and the stage between laboratory proof and commercial use is underfunded. Value arrives late and work is done twice.
Fund the gap, and move together. Testing research against real use produces evidence that benefits a whole field, so a single private investor rarely captures enough of the return to fund it alone. Public and pooled ecosystem money fills that gap and makes later private investment easier to justify. Seat researchers in the build with data access and the right to publish, so that each side’s evidence is checked by the other. Where a market is locked, have one actor from each side commit to one small case, on condition that the others do.
Act on the evidence you have before the system is ready to back you, at a scale where being wrong is affordable.
The blockers share a root: an actor optimising its own position within the current system, at a cost to the system’s capacity to change. That makes this less a cultural problem than one of structure and incentive design.
the.initiator
Design rules are not adopted by markets. They are adopted by one actor, and then by others. The initiator is the actor that makes the shifts before its market rewards them. It can be a founder, a company, a fund, a regulator or a community, and in each case it acts through people who decide. Our proposed sequence is that one actor changes what it optimises for, its behaviour changes accordingly, others with compatible interests join, and structure follows to protect what they have built.
Why going first can be rational.
This manifesto has argued that actors follow their incentives, so the initiator cannot be the exception. It does not act against its own interest. It changes which interest it pursues, and three mechanisms can make that pay.
What it costs.
The first actor usually pays more than those who follow. It funds the first builds, absorbs the first failures and carries the risk that nobody joins. The proposition holds only where the three gains can outweigh that cost, which is why it is limited to markets with lock-in and network dependence.
Where it stops.
If the initiator’s advantage comes to depend on controlling the shared layer, the incentive to close it returns. Ecosystem value first, own position second, works because the second is made to depend on the first. The contestability rules of shift 5 are what keep it that way.
The six first moves in section 02 are where an initiator begins. None of them needs permission. Each needs one actor for whom the calculation works, and who acts on it.
the.flywheel
The initiator’s first move fades unless others can join it. The flywheel is our model of how they do: an inclusive ecosystem on shared, contestable infrastructure that incumbents and newcomers can both use, built up through four recurring stages.
The initiator starts the loop. Read clockwise from stage 1. Each turn should make the next one cheaper.
A working solution does not distribute itself, and those who build are rarely those who operate. Start with the minimum complete set: the smallest group of roles that can take one thing from built to used. In a locked market that means one actor from each side of the transaction and, where research or regulation matters, the researchers and the rule-makers too. Hold roles by function and not by ownership, so that they can change hands as the market matures.
Draws on shifts 4 and 6.Let partners join at any time, avoid vertical lock-in, keep every component replaceable, and let partners specialise. The replacement right has to cover the orchestrator, and it is credible only if a component is actually swapped from time to time. Keep entry open to actors the core does not yet serve, because they are often the first to see what the core’s metrics miss.
Draws on shifts 5 and 3.Build toward an immediate objective with clear value, and start small. Write down the thesis and the stop criteria before each build. Seek proof from outside the group, and pay on proof delivered, whether the answer is yes or no. Research partners co-build, so that the build is itself the test between laboratory and market, and the partners who will reuse the result fund it together. Then extend what works to neighbouring cases.
Draws on shifts 1 and 2.Turn what was built into capabilities that others can reuse and combine. Make bridges, adapters, data export and shared compliance the first of them, so that joining does not require leaving the old system at once and compliance cost stops working as an entry barrier. When a build stops, harvest its parts. Publish what was learned, failures included, and credit whoever produced it, so that being reused builds reputation.
Draws on shifts 5 and 2.Inclusive ecosystem, open to incumbents and newcomers alike. What stage 4 produces is meant to lower the cost of the next turn.
The core proposition.
When both sides can join shared infrastructure, joining costs less than defending, and the side that is losing stops sooner.
The proposition is that shared infrastructure can lower the cost of joining with each turn, while the cost of defending a rival system rises as others join. At some point joining is the cheaper option even for those who would have preferred to win.
Each term needs a definition.
- Joining cost
- What an actor pays to take part: integration work, the data and practices it has to share, and the control it gives up.
- Defending cost
- What it pays to hold out: maintaining a rival standard or a closed system, the customers and partners it loses while the market stays split, and the risk of being stranded on the losing side.
- Open
- Contestable in the sense of shift 5: interoperable, portable, replaceable, with a credible exit, under governance that resists capture.
- Transition duration
- The time from a technology being proven to its broad use. Broad use means it has become the default for most of the relevant market, not only for early adopters.
Why each turn should be cheaper.
Reusable components mean a later actor integrates what an earlier one built. Shared evidence means it does not repeat the earlier one’s failed experiments. Interoperability means its existing systems keep working. This is the network effect we claim for the shifts: each actor that makes them lowers the cost for the next.
The counterweight.
More participants also mean more coordination: more interfaces to agree, slower decisions, more governance. The flywheel works only while the gains from reuse, shared evidence and interoperability exceed the additional coordination costs. Where they do not, a smaller or more tightly governed group will move faster, and should.
One scoreboard per turn. Roles filled, partners joined, proofs gained, components reused, kept apart from each partner’s core metrics.
Where to enter. With the minimum complete set and one small build. Completeness is a direction, not a condition for starting.
What keeps it contestable. A flywheel that succeeds is tempted to close, and the network effects that make it attractive are what would let a winner lock others in. The rules of shift 5 therefore apply to the flywheel itself, its orchestrator included.
What it does not do.
It builds around a locked market and does not unlock it. It does not move an incumbent that already has the regulator on its side. Reputation inside the ecosystem does not change who is promoted inside each partner. Where it works, it works by making the prevailing model unnecessary.
coda
“A technology’s value reaches most people only when the infrastructure it runs on stays contestable, and when the actors inside it keep making the decisions that hold it open. Neither happens by default. You are one of those actors.”
This manifesto is an invitation to builders, investors and institutions to treat the next transition as a decision about incentives and shared infrastructure, and to test this model with us.