An App Is Not a Wrapper Around Your Site: Where the Line Falls

“We need an app — the same as our site, only in the store” comes up at every other meeting. Sometimes it is justified; more often it is not. The user spends storage, sits through an install and grants permissions; in return they expect something the browser cannot offer. When that something is missing, the app is deleted within a week — and the budget has already gone on two platforms and their upkeep.

1. When an App Is Genuinely Needed

Frequency — the product is used several times a week: delivery, banking, logistics, internal systems for staff.

Offline — a courier, an installer or a driver works where the signal drops; data has to accumulate locally and sync later.

Hardware — a camera with recognition, a barcode scanner, background GPS, Bluetooth, NFC.

Push notifications as part of the process — not a newsletter, but a signal without which the flow does not work: a new order, a status change, a confirmation.

Speed and the feel of response — interfaces with frequent data entry, where every extra page reload costs an employee time.

If none of these apply, a responsive site or a PWA will close the task more cheaply and faster.

An app is justified where it gives what the browser cannot. Everything else is the cost of two distributions instead of one.

2. Native or Cross-Platform

For most business tasks, cross-platform development covers 90% of the functionality from a single codebase. A native implementation stays justified where complex graphics, real-time camera work, background processes with hard requirements, or deep integration with the platform's system capabilities are critical. The practical compromise is a shared codebase with native modules for the bottlenecks.

3. What Estimates Leave Out

The cost of an app does not end at release. The budget has to carry annual developer accounts, updates for new OS versions, store review (including resubmissions), support for old versions on the devices of users who never update, and crash reporting. That is usually 15–25% of the development cost per year — a line item better seen before the start than a year in.

4. A Practical Rule

Start with the web if the scenario allows it, and collect data: how many users arrive from mobile, how often, which actions repeat. An app built on that data hits the target far more accurately than an app built on assumptions.