Mobile

Designing Apps for the Phones Your Users Actually Own

Every app we publish is tested on an entry-level Android handset, not the newest phone in the office. It changes almost every decision, and it is the single practice that most improves how an app performs in the field.

  • KeyCode Team
  • 24 May 2026
  • 1 min read

Assume the network will disappear mid-action

A surveyor in a village does not get a spinner and a retry button. Writes go to local storage first and sync when a signal returns, with conflicts resolved by a rule agreed in advance rather than by whichever request landed last.

Budget memory like it costs money

On a 2 GB device the operating system will kill a greedy app without warning, and the user will read that as "the app crashed".

  • Load lists in pages; never hold a full dataset in memory.
  • Downscale images at capture, not at upload.
  • Release map and camera resources the moment the screen is left.

Design for one thumb in bad light

Larger tap targets, higher contrast, and the primary action within thumb reach at the bottom of the screen. These are not accessibility extras — they are what makes an app usable standing in a shop doorway at four in the afternoon.

The takeaway

Test on the cheapest phone your users are likely to own. Everything that follows from that constraint makes the app better on the expensive ones too.

Want this applied to your business?

Tell us what you are trying to fix. We will come back with a scope, a timeline and a fixed price — usually within one working day.