The honest answer is that the range is wide because the word “app” covers everything from a styled content reader to a regulated financial product.
What actually drives the number
Complexity, first and by a distance. A list-and-detail app with a content backend is a different engineering problem from one with payments, real-time sync, offline conflict resolution or background location. Feature count matters less than whether any single feature is genuinely hard.
Integration surface. Every external system your app has to talk to — a payment processor, a CRM, a legacy internal database, an identity provider — adds work that is invisible in a feature list but real in a schedule.
Regulatory requirements. Health and financial data change what “done” means. The work is not just more features, it is a higher bar for every feature.
Platform count matters less than people expect. Going from one platform to two on a cross-platform build is not a doubling; it is usually a modest increment. Choosing native for both platforms, however, genuinely is close to two builds.
What we do
We scope before quoting, because a number produced before understanding the integration surface is a guess that one party will end up absorbing.
Ongoing maintenance typically runs 15-20% of the initial build per year. Budget for it — an app that stops being updated stops working, usually at the next September OS release.