Building for the 2.6 billion people still offline

In 2024, according to the International Telecommunication Union, 5.5 billion¹ people were online. That is a remarkable achievement for a technology most of the world did not use a generation ago. The same report counts 2.6 billion² people, about a third of the world, who were still offline. In low-income countries, only 27% of people use the internet².

Those two numbers are worth holding together. Most people now use the internet, and a very large number still do not. Between them sits a group technology companies rarely design for: people who are online some of the time, on a shared phone, on a slow or costly connection, in places where the signal comes and goes.

Connected is not a yes or no

Much software is built on an unspoken assumption: the user has a recent phone or laptop, a fast connection that does not drop, and a data plan that does not run out. For a designer in a city office, this is simply normal. For a shopkeeper in a market town, a community health worker on a rural route or a student sharing the family phone, it is not.

When software assumes a perfect connection, it fails in ways that feel personal. The form that loses everything when the signal drops. The page that shows a spinner forever. The app that needs an update larger than the month’s data allowance. Each of these tells the user, quietly, that the product was not made for them.

A spinner that never stops is a small way of telling someone the product was not made for them.

Designing for the edge

Building for patchy connections is less glamorous than building for fast ones, and it is harder. It means making pages small, so they arrive quickly on slow networks. It means keeping what has already been loaded, so a lost connection does not mean a lost screen. It means text that is readable on small, dim screens in bright sunlight, and buttons large enough for a thumb.

It also means pricing that does not assume a steady income. People at the edge of connectivity are often at the edge of the formal economy too. Upfront annual fees and card-only sign-ups shut them out before the first page loads.

And it means patience in the design of every screen. A person on a slow connection should never be punished for pressing a button twice. A form should not throw away what someone typed because a request timed out. An error message should say what happened and what to do next, in plain words, rather than a code. These are small courtesies. Added together, they decide whether software feels built for you or merely tolerant of you.

The people who are not online yet

For the 2.6 billion who are fully offline², software is not the first problem. Coverage, device cost, electricity, literacy and relevance all come first, and they are mostly the work of governments, network operators and civil society. We do not pretend that a better app brings anyone online.

What software companies can do is make sure that when people do arrive, what they find works for them. And many people who are offline are already affected by software indirectly. The shop they buy from, the clinic they visit and the school their children attend increasingly run on it. If those systems fail when the connection drops, the offline customer is the one left waiting at the counter.

Put differently, building for unreliable connections is a way of serving people who may never open our software themselves. The receptionist who can still read the day’s bookings when the network fails is protecting the patient who has walked an hour to be there. That is reason enough to take the edge seriously.

It is also a matter of fairness between customers. A business that runs well only when the network is perfect is quietly offering a better service to people in well-connected places. Software that copes with gaps narrows that difference, even if only a little.

What we have built

Svayam Incarnation Limited designs Intersink with unreliable connections in mind. Pages you have already opened keep working when you go offline, so a receptionist can still read today’s bookings and a shop can still check its price list. Avatars are drawn on your device instead of downloaded, which keeps pages lighter.

Pricing follows the same thinking. The User ID is free. Businesses install only the apps they need, and pay after the month is used. Students, small businesses, startups and nonprofits can apply for free credits, which are used before any card is charged.

This is not a complete answer, and there is a lot more to do. Offline reading covers pages you have already opened; it is not full offline working for every task. We say so plainly, because being honest about limits is part of designing for people who cannot afford to be let down.

Build for the slowest connection your customers use, and everyone else gets a faster product.

There is a practical bonus in all this. Software that works on a weak signal also works better on a strong one. Pages that are small for a village are fast in a city. Screens that are clear on a cheap phone are clear on an expensive one. Designing for the edge rarely costs the centre anything.

The ITU’s figures will improve, and we hope quickly. In the meantime, the people who are online some of the time deserve software that treats their connection as normal, not as an error to be apologised for.