Building For Low Bandwidth: The African Product Playbook

The most common product mistake on the continent is treating a slow connection as an edge case. For most users it is the default condition, and it shapes every decision that follows.

Design for cost, not just speed

Users track data spend the way they track airtime. An app that silently pulls ten megabytes of video on open gets uninstalled — not because it is slow, but because it is expensive. Autoplay off by default, aggressive image compression, and an honest “this will use X MB” before a heavy action all increase retention.

Assume interruption

Sessions end mid-action. Save state locally and continuously, make every form resumable, queue writes and sync when the connection returns. An app that loses a half-finished upload has lost the user.

One device, several people

Phones are shared across a household or a shop. Fast account switching, a lightweight guest mode and no assumption that the last session belongs to the current person are basic requirements, not privacy extras.

Optimise for the second-hand Android on a 3G cell, and the flagship on fibre takes care of itself.

Payments are the real product

Card penetration is low and card abandonment is high. Bank transfer, USSD, mobile money and agent networks are not fallbacks — for large parts of the market they are the primary rail. Price in local currency, in amounts that match how people actually spend: small, frequent, immediate.

The performance budget that matters

Under 200KB for a first meaningful screen. Server-rendered HTML before hydration. Fonts subset or dropped entirely. Lazy-load anything below the fold. These are not micro-optimisations in this market; they are the difference between a product that opens and one that does not.