Why Digital Platforms Fail the World's Poorest

Why Digital Platforms Fail the World’s Poorest

The promise was simple: give a poor farmer in Kenya a smartphone, connect her to a digital platform, and she would leapfrog into the modern economy. She could sell crops directly to buyers, skip exploitative middlemen, and access credit based on her transaction history. It was a beautiful story. And for a tiny fraction of users, it actually worked.

For everyone else, the platform economy has been less a ladder and more a trap.

A 2021 paper by Carla Bonina at the University of Surrey, Kari Koskinen and Ben Eaton at the University of Copenhagen, and Annabelle Gawer at Imperial College London dug into why. Their review of the existing research on digital platforms in the global South found something uncomfortable: the very features that make platforms so powerful in rich countries often become liabilities in poor ones. The authors call for a new research agenda on what they term “digital platforms for development,” but their diagnosis of what is broken is already stark (Bonina et al., 2021).

The problem is not that the technology does not work. It is that the business models do not fit.

The Two Kinds of Platforms, and Why Only One Gets Built

Bonina and her colleagues start by drawing a clean distinction that most tech enthusiasts miss. There are two types of digital platforms, and they serve very different purposes.

Transaction platforms are what most people think of when they hear the word “platform.” Uber, Airbnb, Amazon Marketplace. They connect buyers and sellers, take a cut, and optimize for efficient exchange. Innovation platforms are different. They provide a foundation upon which others can build their own products. Think of the Android operating system, or the cloud computing services that let startups launch without owning servers.

The authors note that transaction platforms have received the vast majority of attention and investment in the global South. Ride-hailing apps, e-commerce marketplaces, freelance work platforms. These are the tools that development organizations and venture capitalists love. They promise immediate, measurable impact.

But Bonina et al. found that innovation platforms are arguably more important for long term development. They allow local entrepreneurs to create solutions tailored to their own communities. A farmer in rural India does not need a generic marketplace. She needs a platform that understands local languages, seasonal weather patterns, and the informal credit systems that actually govern her life. Innovation platforms could enable local developers to build that. Instead, most of the global South is being flooded with transaction platforms designed in San Francisco or Shenzhen, optimized for users with bank accounts, stable addresses, and reliable internet.

The result is a mismatch that the authors describe as a fundamental tension between platform business models and developmental goals.

Why Free Markets Are Not Free for the Poor

Here is where the research gets uncomfortable. Bonina and her team synthesized dozens of studies and identified several recurring problems. One of the most troubling is what they call the exacerbation of inequalities.

The logic of platform economics rewards scale. Network effects mean that the biggest platform wins. In a wealthy country with good infrastructure, that might mean better service and lower prices. In a poor country, it often means something else: a foreign owned platform extracts value while local competitors cannot get started.

The authors point to evidence that digital platforms in the global South frequently reinforce existing power structures rather than disrupting them. The middlemen that platforms were supposed to eliminate? They often adapt and become platform operators themselves, using their local knowledge and capital to control access. The poor farmer ends up paying fees to a different set of intermediaries, with the added cost of the platform’s commission.

Bonina et al. also flag a problem that rarely gets discussed in Silicon Valley: the dark side of platforms. This includes algorithmic bias, labor exploitation, and the erosion of local institutions. When a platform replaces a traditional market, it does not just change how transactions happen. It changes who has power, who sets rules, and who gets left out.

One study they reviewed found that digital platforms in Africa often require users to have a smartphone, a bank account, and a formal address. That excludes the majority of the population. The people who need platforms the most are precisely the ones who cannot access them.

The Institutional Trap

The most interesting argument in the paper is about institutions. Bonina and her colleagues argue that platforms do not operate in a vacuum. They interact with existing laws, norms, and power structures. In the global South, those institutions are often weak, informal, or corrupt.

A platform that works well in Denmark, where contracts are enforced and trust is high, may fail catastrophically in a place where disputes are settled by village elders and property rights are unclear. The platform cannot simply replace those institutions. It has to work with them. But most platform designs assume that formal institutions exist and function.

The authors raise a provocative question: what happens when platforms undermine the very institutions that poor communities rely on? A digital lending platform might offer quick credit, but if it does not respect local repayment norms or customary debt forgiveness practices, it can destroy social trust. The platform becomes a source of conflict rather than a tool for development.

Bonina et al. call for research on how platforms can be designed to strengthen, rather than weaken, local institutions. But they do not pretend this is easy. It requires platform builders to understand contexts that are radically different from their own.

What the Research Does Not Prove

This paper is a review, not an experiment. It synthesizes existing studies and identifies gaps. It does not claim to have proven that digital platforms always harm the poor. In fact, the authors are careful to note that there are cases where platforms have helped. Mobile money in East Africa is the most famous example. M Pesa, a transaction platform, genuinely increased financial inclusion for millions of people.

The open question is why some platforms succeed while most fail. Bonina and her team suspect the answer lies in the specific characteristics of each platform and its context. A transaction platform that builds on existing informal practices, like M Pesa did with airtime transfer, might work. One that tries to impose a foreign model probably will not.

The authors also do not address the question of whether platforms can ever be truly developmental if they are owned by foreign corporations. This is a political question as much as a technical one, and they leave it for future research.

The Indigenous Innovation Gap

One of the most promising directions the authors identify is indigenous innovation. They define this as platforms built by local entrepreneurs, for local needs, using local knowledge. This is not just about translation or superficial localization. It is about building entirely different systems that reflect different values and priorities.

An indigenous platform might prioritize community resilience over individual profit. It might value long term relationships over short term transaction volume. It might use voice interfaces instead of text, because literacy rates are low. It might accept payments in kind, because cash is scarce.

Bonina et al. found almost no research on this topic. The platform literature is dominated by studies of Western platforms expanding into new markets. There is very little about platforms that emerge from within poor communities, built by and for the people who live there.

This is not an accident. Venture capital does not fund platforms that cannot scale to billions of users. Development aid does not fund projects that cannot show measurable impact in two years. The incentives are all wrong.

What This Actually Means

• ▸Stop funding generic platforms. Development organizations and impact investors should stop pouring money into clones of Uber or Airbnb for poor countries. The evidence suggests these rarely work and often cause harm. Instead, fund platforms that are designed for specific local contexts, with local partners who understand the institutions and norms.

• ▸Measure exclusion, not just inclusion. Every platform report brags about how many users it has. Almost none report how many people tried to use the platform and could not. Bonina et al. make clear that the barriers to access are not just technical. They are social, economic, and institutional. Platforms should be required to measure and report their exclusion rates.

• ▸Build innovation platforms, not just transaction platforms. The most valuable platforms for development may be the ones that enable others to build, not the ones that extract fees from transactions. Governments and donors should invest in open infrastructure that local entrepreneurs can build on, rather than trying to pick winners.

• ▸Design for weak institutions, not strong ones. Platforms intended for the global South should assume that formal contract enforcement, reliable identity systems, and functioning courts do not exist. They should be designed to work with informal institutions, not against them. This is harder, but it is the only way to reach the people who need help most.

• ▸Let the poor define value. The authors call for research on alternative forms of value. Development should not mean adopting Silicon Valley metrics. A platform that helps a community manage shared resources, resolve disputes, or preserve cultural knowledge may be more valuable than one that increases transaction volume. Let the users decide what counts as success.

The digital platform economy is not going away. But the fantasy that it will automatically lift the world’s poor is dead. The research is clear: without fundamental redesign, platforms will continue to fail the people who need them most. The question is whether anyone is willing to build something different.