The honest starting point is that most businesses should not build an app. Apps earn their cost only when purchase frequency is high enough that customers want a faster path than the browser.
When an app is justified
Frequent repeat purchase such as grocery, food and pharmacy; field or driver operations that need offline capability and scanning; and booking relationships where the customer returns regularly. Outside these, a fast mobile site usually delivers more for less.
Engagement comes from utility, not notifications
Saved lists, one-tap reorder, order tracking and stored preferences create genuine reasons to open the app. Push notifications sustain a habit that already exists; they cannot create one, and over-used they drive uninstalls.
Android matters more here than global figures suggest
The UAE's resident population skews Android more heavily than headline regional statistics imply, particularly in the segments that buy groceries, remit money and use delivery services most frequently. Building iOS first on instinct can miss a large part of the actual customer base.
Measure retention, not installs
Installs are a marketing cost. The numbers that matter are thirty-day retention, orders per active user and the difference in repeat rate between app users and web users. If that difference is not material, the app is not paying for itself.