Centralize the Truth, Decentralize the Power (Part 1): The Kampong, the Platform, and Who Owns What We Depend On

A two-part series on business philosophy for the age of platforms and AI. Part 1 explores the ideas. Part 2 turns them into decisions you can make on Monday morning.

---

Should a business depend on one big platform, or on a network it partly owns? The short answer: centralize the truth (your data, your records, your single view of the business) but decentralize the power (who makes decisions, who captures the value, who can switch you off). Businesses and communities that confuse the two either become fragile or become someone else’s tenant.

That sentence took us more than twenty years of implementing systems for Singapore SMEs to be able to write. This article explains where it comes from, using three thinkers most business owners have never read, one Nobel Prize that was split between two opposite ideas, and a very old Malay word.

---

What can a kampong teach a modern business?

Picture a kampong in 1950s Singapore. Timber houses on stilts, zinc roofs, chickens under the floorboards. Water came from a shared well or standpipe. Nobody owned it the way you own a car. Yet it was rarely broken, rarely abused, and rarely fought over.

When a family needed to move their house, the neighbours did not call a contractor. They showed up, slid poles under the frame, and carried the whole house on their shoulders to its new spot. The Malay word for this is gotong royong: mutual help, carrying the load together.

Now picture a modern Singapore SME in 2026. Orders arrive through a marketplace. Deliveries go through a logistics app. Customers pay through a payment gateway. Marketing runs through an ad platform. Staff chat through a messaging app. The business owns almost none of the roads its revenue travels on.

Both pictures are communities of dependence. Nobody in the kampong was self-sufficient either. The difference is not whether you depend on others. It is who owns the thing you depend on, and whether you have a voice when the rules change.

That is the question this series is about. It is an old question. It just wears new clothes every few decades.

---

Why do platforms become more expensive over time?

Almost every digital platform follows a recognisable life cycle. It is worth describing plainly, without villains, because the pattern is structural, not personal.

Stage 1: The generous phase. A new platform needs two sides to show up: buyers and sellers, riders and drivers, guests and hosts. So it subsidises both. Rides are cheap. Commissions are low. Sign-up bonuses are everywhere. Investors fund the gap.

Stage 2: The dependence phase. Because the platform is convenient, people stop maintaining alternatives. Shops close their own delivery fleet. Customers delete other apps. Drivers stop building their own client lists. The platform becomes the road everyone travels on.

Stage 3: The toll phase. Investors eventually need returns. The platform now controls the road, so it raises the toll. Commissions creep up. New “platform fees” appear. Visibility that used to be free now needs to be bought as advertising.

Writer Cory Doctorow gave this pattern a memorable, if impolite, name that became a word of the year. You do not need the label to recognise it. Every SME owner who has watched a marketplace commission rise, or paid for ads to reach customers who already follow them, has lived it.

To be fair, fee increases are not always greed. In Singapore, the Platform Workers Act now requires ride-hailing and delivery platforms to contribute to their workers’ CPF, and that cost is real. When Grab raised its ride platform fee from 90 cents to S$1.20 from 1 January 2026, it pointed directly to those higher CPF obligations. Costs move through the system, and someone pays.

But notice the deeper point. The person paying has no vote on the fee. The rider cannot propose a different design. The driver cannot audit the calculation. The merchant cannot negotiate as a group. That is not a pricing problem. It is a governance problem.

And governance is where philosophy becomes surprisingly practical.

---

Is the “tragedy of the commons” true?

In 1968, ecologist Garrett Hardin published an essay in the journal Science called “The Tragedy of the Commons.” His argument was elegant and gloomy.

Imagine a pasture open to all herders. Each herder gains the full benefit of adding one more cow, but shares the cost of overgrazing with everyone. So each rational herder keeps adding cows. The pasture is destroyed. Hardin’s conclusion: shared resources are doomed unless they are either privatised (someone owns the pasture) or controlled by the state (a government rations it).

The idea was not new. More than two thousand years earlier, Aristotle observed in his Politics that whatever is shared by the most people tends to receive the least care, because everyone assumes someone else will look after it.

For decades, Hardin’s essay shaped economic policy. And it quietly shaped business thinking too. The logic runs like this: shared things fail, so someone must own the platform. Every centralised marketplace is, in a sense, Hardin’s private owner of the pasture.

There is just one problem. In the real world, many commons did not collapse.

---

What did Elinor Ostrom discover that most business owners have never heard of?

Elinor Ostrom was an American political scientist who did something unusual for an academic of her era: she went and looked.

She studied fishing grounds in Turkey, irrigation systems in Nepal and Spain, mountain pastures in Switzerland, and forests in Japan. Many of these shared resources had been managed by communities, without a private owner and without a central government, for hundreds of years. Some Swiss alpine meadows had community rules dating back to the 1500s.

Her 1990 book, Governing the Commons, showed that Hardin’s two options were not the only options. There was a third: communities that govern themselves, under the right conditions.

In 2009, Ostrom became the first woman to win the Nobel Memorial Prize in Economic Sciences.

Here is the detail that turns this into a business story. She shared that prize with Oliver Williamson, an economist who studied almost the opposite question: why companies choose hierarchy, contracts, and central control instead of open markets. The Nobel committee put them together under one theme: economic governance.

Think about what that means. The most prestigious prize in economics was split between the person who explained when self-governing communities work and the person who explained when central control works. Neither was declared the winner. The lesson is not “decentralization good, centralization bad.” The lesson is that the governance design has to fit the situation.

We will come back to Williamson later, because he deserves a fair hearing.

---

What are Ostrom’s design principles, translated for a modern business?

Ostrom identified a set of design principles shared by long-lasting commons. They were written about fisheries and forests. Read them again as if they were about a platform, a supplier network, or a group of businesses sharing a system. They translate remarkably well.

1. Clear boundaries. Everyone knows who is a member and what the shared resource is. Business translation: A network needs a clear membership list and a clear definition of what is being shared: delivery capacity, customer data, compute, purchasing power.

2. Rules that fit local conditions. The rules for a Swiss meadow are not the rules for a Nepali canal. Business translation: A one-size-fits-all platform policy, set in a distant headquarters, will always fit someone badly. Good networks let local rules vary.

3. The people affected help make the rules. Business translation: This is the principle platforms most often break. Drivers, merchants, and freelancers are the most affected and the least consulted.

4. Monitoring by people accountable to the community. Someone watches, and the watchers answer to the members. Business translation: This is exactly the “watcher” role in modern decentralized network designs. Quality is checked by independent parties, not only by the platform that profits from the transactions.

5. Graduated penalties. A first mistake gets a warning, not an expulsion. Business translation: Compare that with a gig worker deactivated by an algorithm with no explanation and no appeal.

6. Cheap, accessible ways to resolve disputes. Business translation: A merchant should not need a lawyer to challenge a platform decision.

7. Outside authorities respect the community’s right to organise. Business translation: Regulators matter. A network can only self-govern if the law recognises it.

8. Nested layers. Small groups govern local matters; larger groups handle wider issues. Business translation: A neighbourhood delivery group, inside a city federation, inside a national standard. That is how resilient networks scale.

Read those eight principles once more and notice something. None of them require blockchain, tokens, or any technology at all. They are about human design. Technology can make them cheaper to run, but it cannot replace them. A decentralized app with no dispute process and no voice for members is just a centralized platform with extra steps.

---

Why does local knowledge beat central planning?

In 1945, the economist Friedrich Hayek published an essay called “The Use of Knowledge in Society.” His central insight is one every SME owner already knows in their bones, even if they have never read him.

The knowledge needed to run an economy is not sitting in one place. It is scattered in millions of heads, in tiny, local, often unwritten forms. The shopkeeper knows that this customer always pays late but always pays. The driver knows that this loading bay is blocked every Thursday after 3pm. The technician knows that this machine makes a particular sound a week before it fails.

Hayek called this “knowledge of the particular circumstances of time and place.” No central planner can collect it all, because much of it cannot even be written down. It lives in experience.

His argument was aimed at centrally planned economies. But it applies with surprising force to centrally planned platforms. When a single algorithm sets prices, routes, and rankings for an entire city, it is making decisions with only the knowledge it can measure. Everything the driver, the shopkeeper, and the technician know, but the app cannot see, is thrown away.

Here is the uncomfortable part for anyone who sells software (including us). An ERP system can centralize data. It cannot centralize wisdom. The best implementations we have seen do not replace local judgement. They give the people closest to the work better information, and then trust them to decide.

---

What is subsidiarity, and why should a business care?

Subsidiarity sounds like jargon. It is actually one of the most useful management ideas you have probably never been taught.

The principle says: a decision should be made at the lowest level capable of making it well. Higher levels should step in only when lower levels genuinely cannot handle something.

The idea was set out formally in a 1931 Catholic social teaching document, Quadragesimo Anno, which argued that it is wrong for a larger body to take over what a smaller community can do for itself. Decades later, the European Union wrote subsidiarity into its founding treaties, as the rule for deciding when Brussels should act and when member states should.

Now apply it to a business.

• Should the warehouse supervisor need head office approval to re-order packing tape? Subsidiarity says no.
• Should a single store be able to set its own supplier contracts that affect group-wide pricing? Subsidiarity says probably not; that decision belongs a level up.
• Should a delivery platform headquartered in another country decide the pay rules for a driver in Pasir Ris? Subsidiarity asks: who is the lowest level actually capable of deciding this well?

Subsidiarity is not the same as “decentralize everything.” It is a sorting rule. Some decisions really do belong at the top: safety standards, data formats, legal compliance. Others belong at the edge. The mistake is putting everything in one place, whichever place that is.

This becomes important in Part 2, when we look at AI agents. An AI agent is, among other things, a way of pushing decisions down to a very low level. Whether that is empowering or dangerous depends entirely on whether we sort the decisions well.

---

Why do centralized systems break in surprising ways?

In 1956, the cybernetician W. Ross Ashby described what became known as the Law of Requisite Variety. In plain language: a system that controls something must be at least as varied as the thing it controls.

A thermostat with only “on” and “off” cannot manage a room with many kinds of disturbance. A manager with one playbook cannot manage a team facing a hundred different situations. A single platform policy cannot manage a hundred thousand different small businesses.

This is why centralized systems tend to fail in two predictable ways:

• They oversimplify. To manage enormous variety, the centre reduces everything to a few measurable numbers: star ratings, acceptance rates, response times. Anything that does not fit the numbers is invisible.
• They become a single point of failure. When the centre goes down, everything goes down together.

The second failure is not theoretical. The draft whitepaper for the Decentralized Infrastructure Network (DIN), written by the team behind Infura, one of the largest blockchain infrastructure providers, describes exactly this moment. In November 2021, a software bug caused a serious outage at Infura, and prominent applications that depended on it, including the MetaMask wallet, were affected. The team wrote that the incident became a turning point that pushed them to design a network where many independent providers share the load instead of one company carrying it.

That is Ashby’s law in action. One provider, however competent, has limited variety. Many providers, coordinated well, have more.

---

So is centralization always the villain?

No. And any article that tells you otherwise is selling something.

Remember Oliver Williamson, Ostrom’s co-laureate. His work, building on an earlier idea from economist Ronald Coase, asked a simple question: if markets are so efficient, why do companies exist at all? Why don’t we just hire everyone for each task through a market?

The answer is transaction costs. Finding partners, negotiating, checking quality, enforcing agreements: these all cost time and money. When those costs are high, it is cheaper to bring the work inside one organisation and manage it with hierarchy. Central control is not a failure of imagination. It is often a rational response to the cost of coordinating many independent parties.

So let us give centralized platforms their strongest possible case.

They solve the empty-room problem. A marketplace is only useful if the other side is already there. A central platform can spend money to fill both sides at once. A community network often cannot, and dies before it reaches critical mass.

They make trust cheap. When you book a ride from a stranger, you trust the brand, the background checks, the insurance, and the dispute process. Building that trust across a loose network is hard and expensive.

They move fast. One decision-maker can ship a feature in a week. A federation of members may take a year to agree on the colour of the button.

They carry heavy obligations. Insurance, compliance, safety, CPF contributions under Singapore’s Platform Workers Act: these costs do not disappear in a decentralized model. They simply have to be carried by someone else, and small cooperatives often struggle to carry them.

They often did create enormous value. Millions of people earned income, and millions of small businesses reached customers they never could have reached alone, because centralized platforms existed.

A fair-minded business owner should hold both truths at once. Platforms create real value. Platforms also concentrate power over that value. The question is not whether to use them. The question is how much of your business should sit on roads you do not own, and whether there are better roads being built.

---

If centralization and decentralization both work, how do you decide?

Here is the synthesis that the rest of this series builds on. It resolves what looks like a contradiction in our own work.

For years, we have told clients to build a single source of truth: one ERP, one customer record, one version of the numbers. That sounds like centralization. And it is. So how can the same company argue for decentralization?

Because truth and power are different things.

Centralize the truth. Your records should be unified, consistent, and trustworthy. Two departments with two different stock counts is not “decentralization.” It is chaos. Shared standards, shared data formats, and shared records are what make cooperation possible. The kampong well worked partly because everyone agreed on the rules for using it.

Decentralize the power. Decisions should be made as close to the work as possible (subsidiarity). The value created should flow to the people creating it (Ostrom). The ability to switch providers, change rules, or walk away should stay with the participants (the lesson of the platform life cycle). And no single failure should bring the whole system down (Ashby).

Put simply:

	Centralize	Decentralize
Data and records	✓ one trustworthy version
Standards and formats	✓ everyone speaks the same language
Decisions	only what the edge cannot handle	✓ as close to the work as possible
Ownership of value		✓ with those who create it
Ability to exit		✓ with the participants
Points of failure		✓ spread across many

The most interesting systems being built today, from national e-invoicing networks to open ride-hailing protocols to AI agent standards, follow exactly this pattern. They centralize the standard and decentralize everything else. That is the subject of Part 2.

---

Key takeaways from Part 1

• Dependence is not the problem; ownership and voice are. Every business depends on others. The real question is whether you have a say when the rules change.
• Platforms follow a predictable life cycle: generous, then indispensable, then extractive. It is structural, not personal.
• The “tragedy of the commons” is not inevitable. Elinor Ostrom showed that communities can govern shared resources for centuries under the right design principles.
• Local knowledge is real and cannot be fully centralized (Hayek). Good systems inform people close to the work; they do not replace them.
• Sort decisions by level (subsidiarity). Some belong at the top; most belong closer to the edge.
• Centralization has a strong case too (Coase and Williamson). Transaction costs, trust, and speed are real.
• The resolution: centralize the truth, decentralize the power.

---

Frequently asked questions

What does “centralize the truth, decentralize the power” mean?

It means a business should keep one trustworthy, shared version of its data, records, and standards, while spreading decision-making, ownership of value, and the ability to switch providers across the people who do the work. Unified data enables cooperation; concentrated power creates fragility and dependence.

What is the tragedy of the commons?

The tragedy of the commons is an idea popularised by Garrett Hardin in 1968: when a resource is shared, each user has an incentive to overuse it, so it gets destroyed. Elinor Ostrom’s later research showed this outcome is not inevitable, because many communities successfully govern shared resources with clear rules and local monitoring.

Why did Elinor Ostrom win the Nobel Prize?

Elinor Ostrom won the 2009 Nobel Memorial Prize in Economic Sciences for showing how communities can manage shared resources such as fisheries, forests, and irrigation systems without either private ownership or state control. She shared the prize with Oliver Williamson, who studied why firms use hierarchy instead of markets.

What is subsidiarity in business?

Subsidiarity is the principle that a decision should be made at the lowest level capable of making it well. In business, it means empowering the people closest to the work to decide routine matters, while reserving higher-level decisions, such as standards, compliance, and group-wide pricing, for the levels that can handle them properly.

Are centralized platforms bad for small businesses?

Not inherently. Centralized platforms help small businesses reach customers, build trust, and move quickly. The risk is dependence: when a platform controls the channel, it can change fees or rules without the business having a voice. The practical goal is to use platforms without letting them own your whole route to the customer.

---

Coming up in Part 2

In Part 2: Networks, AI Agents, and the Three-Question Test, we move from philosophy to practice:

• The quiet decentralized networks already being built around you, including one the Singapore government is requiring every GST-registered business to join.
• What India’s open commerce network and a zero-commission ride-hailing app revealed about who really holds power.
• Why AI agents could become either your most loyal staff or your most expensive landlord.
• A simple three-question test to decide what in your business should be centralized, and what should not.

---

Mxgsoft has helped Singapore SMEs with digital transformation since 2004. If you are thinking about which systems your business should own, and which it should rent, talk to us.

---

References and further reading

• Hardin, G. (1968). “The Tragedy of the Commons.” Science, 162(3859).
• Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
• Hayek, F. A. (1945). “The Use of Knowledge in Society.” American Economic Review.
• Ashby, W. R. (1956). An Introduction to Cybernetics.
• Coase, R. (1937). “The Nature of the Firm.” Economica.
• Infura and DIN Team (2024). Decentralized Infrastructure Network (DIN) Whitepaper [operational draft].
• Malay Mail (24 Dec 2025). Grab to raise ride platform fee in Singapore, citing CPF contributions.