Designing for the Last Bar of Signal
There's a design assumption baked into most modern software that I've had to unlearn: that the network is there. That it's fast, cheap, and always on. Build between Benin City and Porto Alegre for long enough and you stop believing it.
When you build for places where data costs real money and connectivity comes and goes, you start designing for the last bar of signal — the moment when the connection is weakest, not strongest. That single shift changes almost every decision.
The message has to land, not just send
In a well-connected world, "we sent it" and "they got it" are nearly the same sentence. In a low-bandwidth world they are completely different claims. A push notification needs an installed app, an open data connection, and a user who opens the app. Each of those is a place the message can quietly die.
That's why I lean on channels that degrade gracefully:
- SMS reaches any phone, no app, no data. It's the floor that never drops.
- Plain text and small payloads load on a weak connection where a heavy page never finishes.
- Delivery receipts, where you can get them, turn "sent" into "arrived" — the only number that actually matters.
Small is a feature, not a limitation
A 2-megabyte page isn't "rich." On a metered, intermittent connection it's a toll booth. Every image, every script, every font is something the user pays for in money and patience. Designing small — fewer assets, lighter pages, text that survives compression — isn't austerity. It's respect. It says: I know what this costs you, and I've kept it cheap.
The same goes for resilience. Can the task be finished if the connection drops halfway? Can it be retried without doing damage? Can a non-technical person tell whether it worked? In a low-bandwidth context these aren't edge cases. They're the main case.
Why this matters beyond convenience
It's easy to frame low-bandwidth design as a niche constraint. It isn't. It's how a large share of the world actually comes online — on phones, on mobile data, in bursts. When we design only for the fast, cheap, always-on network, we quietly build for the people who already have the most and leave out the people a school, a clinic, or a public service most needs to reach.
Designing for the last bar of signal is, in the end, just designing for everyone honestly — including the people whose connection is the worst on the day it matters most.