Building for a Billion People

There is a number that changes how you think about everything once you sit with it long enough. One billion people. Not as an abstraction — as an actual operating constraint on every decision you make.

I have spent enough time building things in India to know that scale here is not simply more of something. It is a different category of problem. The distance between designing a product for a thousand users and designing one for a billion is not three orders of magnitude — it is a complete reconception of what design means.

What scale actually changes

When you build for a billion people, the tails become the main event.

In a small user population, the unusual cases — the person with the old phone, the slow connection, the non-standard use case — are edge cases. You can document them, acknowledge them, and proceed. At scale, your edge cases contain more people than most countries have as their total population.

This changes what quality means. A 99% reliability rate sounds excellent until you realize that at a billion users, 1% means ten million people experiencing failure simultaneously. The definition of "good enough" has to shift dramatically because the absolute size of your failure is always enormous.

It also changes what diversity means. A billion Indians are not a monolithic group. They speak hundreds of languages across dozens of dialect families. They live in conditions ranging from high-rise Mumbai apartments to villages with four hours of electricity per day. Their relationship with smartphones ranges from digital native to first-time user. Any product that treats this population as uniform will fail significant portions of it, and failure at this scale is not a rounding error — it is an indictment.

The infrastructure prerequisite

Most things that work for a billion people in India work because of infrastructure decisions made years or decades earlier, often by people who never anticipated this specific application.

The mobile network buildout of the 2000s made smartphone access possible at mass scale. The Jio disruption that drove data costs to near-zero is what converted that hardware access into genuine digital participation. The 2016 Jio disruption that drove data costs to near-zero made digital services accessible to people who had never before thought of data as something they could afford to use freely. Aadhaar created a population-scale identity infrastructure that enabled a generation of fintech and public services built on verified identity. UPI built payment infrastructure that has enabled transaction volumes that most global payment systems struggle to match.

None of these were built by startups moving fast and breaking things. They were patient infrastructure investments, often with long incubation periods before the downstream applications became possible. The lesson is that building for a billion requires thinking in infrastructure layers — and accepting that your infrastructure investment might not produce visible returns for years.

What it teaches you

Building for this scale teaches a specific kind of humility that is hard to acquire otherwise.

You learn that your assumptions about users — what they know, what they have, what they want — are almost always wrong in ways that your environment has hidden from you. The team building the product typically lives in conditions that are not representative of the conditions of the people using it. The gap is enormous and largely invisible unless you work actively to close it.

You learn that constraints produce innovation in specific, non-obvious ways. The bandwidth constraints that affected Indian mobile development produced a generation of engineers who understand data efficiency at a level that engineers building for broadband contexts rarely need to develop. Those engineers then applied those instincts to contexts where they produced disproportionate value.

And you learn that scale itself is a form of credibility. When something works for a billion people — genuinely works, across the diversity and complexity of that population — it demonstrates a kind of robustness that no other test can replicate. India is the hardest environment in the world to build for at scale. The things that survive it have usually earned their survival.

What comes next

The next generation of things built for a billion Indians will run on infrastructure that is now in place but not yet fully utilized: cheap data, near-universal smartphone access, population-scale identity verification, real-time payment rails.

What gets built on that infrastructure over the next decade will be among the most interesting and consequential technology development happening anywhere in the world. Not because India is exotic or exceptional, but because the combination of scale, diversity, constraint, and ambition produces problems worth solving and solutions that matter.

If you can build something that works here, you can build it anywhere.