The Hidden Decisions Behind Great Companies

“Olga, I just finished listening to the episode and reading your article.

What a treat!

You did an amazing job understanding the Mercado Libre story

in depth and conveying it to your audience.”

- Stelleo Tolda, co-founded and Board Member of Mercado Libre

 This Week’s Essay

Ecosystems have become one of the most overused ideas in technology.

Platform strategy. Land-and-expand. Multi-product flywheels. It all sounds so deliberate.

But a closer look at the companies that have actually built enduring ecosystems tells a different story. In many of the most successful cases, ecosystems were not engineered upfront.

They emerged when a single product ran up against the limits around it — the trust it could not guarantee, the logistics it could not control, the payments it could not process, the infrastructure it could no longer afford to outsource.

In other words, ecosystems are rarely built out of ambition alone. More often, they are built out of necessity.

Few companies illustrate this more clearly than MercadoLibre.

When Marcos Galperin, Stelleo Tolda, and Hernán Kazah launched MercadoLibre in 1999, the ambition was not to build an ecosystem. It was to build a marketplace — in a region where almost none of the underlying conditions existed. Internet penetration was just 2–3%. There were no established behaviors around buying online. Infrastructure was limited, trust was low, and competition was intense, with more than 80 companies pursuing similar ideas at the same time.

It was, by any definition, a zero-to-one environment.

The initial challenge was not expansion. It was winning the marketplace business.

And winning, in this case, depended on something deceptively simple: getting a transaction to work.

That turned out to be far more complex than matching buyers and sellers.

Early on, MercadoLibre encountered the most fundamental obstacle of digital commerce: trust. Buyers were reluctant to send money to unknown sellers. Sellers were unsure whether buyers would follow through. Transactions broke down not because of lack of demand, but because of lack of confidence.

The company’s response was not to broaden its ambitions, but to intervene directly in the weakest point of the experience. MercadoPago, in its first iteration, was not conceived as a fintech business. It was an escrow layer. The buyer would pay MercadoLibre, the seller would ship the product, and only once the buyer confirmed receipt would the funds be released.

It was a narrow solution to a specific problem.

But it changed the outcome.

Over time, payments evolved into a business of their own. As MercadoLibre connected more payment methods and expanded beyond its own platform, MercadoPago became one of the most powerful fintech platforms in the region. What began as a mechanism to enable trust became an independent engine of growth.

To fully appreciate the scale of what that evolution has produced, just look at the numbers.

Today, Mercado Pago processes close to $300 billion in annual payment volume, serves more than 70 million active users, and manages tens of billions of dollars across payments, credit, and investment products. What began as a simple escrow feature inside a marketplace has become one of the largest financial platforms in the world — and, in many parts of Latin America, a primary interface to the financial system.

That trajectory — from internal fix to standalone platform — is one of the defining patterns of successful ecosystems. But it is only one side of the story.

An equally important set of decisions involves what not to separate.

In the early years, MercadoLibre operated as an asset-light marketplace. It connected buyers and sellers but did not handle the physical movement of goods. Logistics, in theory, could be left to third parties. The company could remain focused on software, preserving margins and avoiding capital intensity.

But that model imposed limits. Delivery was inconsistent. Shipping times were unpredictable. The quality of the end-to-end experience depended on actors MercadoLibre did not control.

In markets like Brazil, where infrastructure was still developing, those constraints were even more pronounced. Growth was not just a function of demand — it was constrained by everything that sat around the product.

At a certain scale, that trade-off became untenable.

The company made a decision that, at the time, ran counter to how it was perceived in the market. It began investing heavily in logistics — warehouses, routing systems, transportation networks. In the short term, this compressed margins and raised questions among investors. In the long term, it transformed the business.

Over time, the share of packages handled within MercadoLibre’s own network grew from a small fraction to roughly 95%, making it one of the largest logistics operations in the region.

Logistics could have become a separate company. It had the scale and the operational complexity to stand on its own.

But separating it would have weakened the system.

Delivery speed, reliability, and cost are not peripheral features in a marketplace. They are central to conversion, retention, and trust. By keeping logistics tightly integrated, MercadoLibre did not simply add another business line. It strengthened the core loop of its marketplace.

This distinction — between what can be expanded independently and what must remain integrated — sits at the heart of ecosystem building. And today, it is becoming existential.

A new wave of Chinese competitors is entering Latin America with the ability to blitzscale at a pace few local companies can match. At the same time, AI is compressing the time it takes to build, ship, and iterate on products. Entire layers of businesses that once took years to build can now be replicated in months.

But logistics does not move at the speed of code. It is physical. It is operational. It requires time, capital, density, and relentless execution. It is not something you can spin up overnight.

It is what, in Jeff Bezos’s words, makes MercadoLibre’s beer taste better.

Amazon offers a parallel case. It began with a single category and, over time, encountered many of the same constraints — fulfillment, reliability, infrastructure — and responded by internalizing them.

Today, Amazon operates one of the largest logistics networks in the world, delivering billions of packages annually and surpassing traditional carriers in scale. Its Prime program, with more than 200 million members globally, is built on that infrastructure — turning logistics into the product.

What was once a backend function has become the experience.

At the same time, Amazon made a different decision with its internal infrastructure. The tools it built to run its own operations — computing, storage, databases — evolved into Amazon Web Services. Unlike logistics, this layer did not depend on the retail experience to create value. It had broad applicability across industries. As a result, it was externalized and scaled as a standalone platform.

The contrast is instructive. Some capabilities become more valuable when they are embedded within a system. Others become more valuable when they are exposed to the outside world.

Across these cases, a pattern begins to emerge.

Ecosystems are not defined by the number of products a company offers. They are defined by how those products relate to one another.

The most successful companies expand along the fault lines of their own experience. They identify the points where the system breaks — where trust erodes, where speed degrades, where coordination fails — and they move to fix them.

Each new layer is not an abstraction, but a response to a specific point of failure.

Over time, those decisions accumulate.

A marketplace becomes a payments company. A payments company becomes a financial platform. Infrastructure built for internal use becomes a global service.

What appears, in hindsight, as a coherent ecosystem is often the result of a series of local decisions, each grounded in the same principle: reduce friction, improve the experience, and bring critical dependencies under control.

This is why attempts to “build an ecosystem” directly so often fall short. They start with the structure rather than the problem. They add products without a clear understanding of how those products reinforce one another. The result is breadth without depth — a collection of offerings rather than a system.

The companies that succeed take a different path.

They begin with a single product, but they do not treat it as a boundary. They treat it as an entry point into a broader experience. And as that experience expands, they make a series of choices — sometimes to extend outward, sometimes to pull capabilities inward, and often to keep key components tightly connected even when separation might seem more attractive in the short term.

What sets them apart is not just strategy. It is time horizon.

They are not optimizing for the next product release or the next earnings call. They are building for a position that will be difficult to displace years — and often decades — from now.

 This Week’s Number

 $87.6B

MercadoLibre’s market capitalization today. When it went public in August 2007 at $18 per share, it was valued at under $800 million. In today’s dollars, that’s roughly $1.2 billion — meaning the company has grown approximately 71x in real terms. Built from a garage in Buenos Aires, in a region with 2–3% internet penetration, against 80 competitors. The ecosystem was never the plan. It was the result.

Community Picks

Five things Stelleo Tolda said in our conversation that we haven’t stopped thinking about.

1. On winning

“It wasn’t about the fees and commissions. It was about the experience, the technology, the trust.”

2. On competition

“Amazon proved to be somewhat slow to the market — which was to our benefit. Brazil was our largest market, a must-win market for us. But not for them.”

3. On building to last

“We have this idea that we’re always in continuous beta. We’re never ready — because we’re always improving.”

4. On AI

“It feels somewhat like it did back then. There’s a lot of effervescence, a lot of excitement. There’s probably some hubris, some valuations that are extremely high. A lot of building of infrastructure — which is similar to what we saw.”

5. On founder transitions

“Sometimes our identity can be wrapped up in what we do. Leaving that abruptly can be hard. In my case it wasn’t — because I was ready for the change.”

Here’s the highlight reel with my favourite moments from the conversation with Stelleo.

What I’m Loving 

Fair warning: this week got away from us. There are five recommendations below. We regret nothing.

1. The Everything Store — Brad Stone

The definitive account of how Amazon went from online bookstore to everything company. Reads differently after this essay — every logistics bet, every infrastructure decision, every sacred cow killed maps directly to what Stelleo described. Required reading for anyone who wants to understand how ecosystems actually get built.

Read it here

2. Brian Chesky’s New Playbook — Lenny’s Podcast

Chesky on killing features, rebuilding from scratch, and why the best product decisions look like retreat before they look like strategy. The Airbnb post-COVID rebuild is one of the clearest recent examples of a company finding its core loop again — directly relevant to the essay’s thesis.

Listen here

3. Patrick & John Collison: A Business State of Mind — Invest Like the Best

Stripe’s founders on why the most important infrastructure is always invisible until it isn’t. Built to solve one specific problem — accepting payments on the internet — and earned every adjacency from there. The MercadoPago parallel is hard to miss.

Listen here

4. John Collison: Growing the Internet Economy — Invest Like the Best

An earlier, equally sharp episode with just John — on conglomerates, the internet economy, and why studying companies that came before you is non-negotiable. More foundational than the one above, and arguably more timeless.

Listen here

5. MercadoLibre Q4 2025 Earnings Call

Ariel Szarfsztejn’s first full year as CEO of one of the most complex tech ecosystems in the world. Watch it as a case study in how a company narrates its own story to the market. Context hits differently after this episode.

Watch here

Thanks for reading,

Olga 

 P.S. If this issue was valuable to you please share it with a founder who needs to hear it. Let’s build LATAM’s next tech leaders—together

🎙 The J Curve  is where LATAM's boldest founders & investors come to talk real strategy, opportunity and leadership.