What a mobile app really costs after you launch it

An Apple developer account costs 99 USD a year, a Google Play account 25 USD once (Apple, 2026; Google Play Console Help, 2026). Those are the only two simple numbers in this subject. Everything else an app costs arrives after it goes live, does not depend on your roadmap, and applies whether or not anyone opens it.

What you pay even if nobody opens the app

Two fees exist regardless of usage. Apple Developer Program membership is 99 USD per year and renews — it is not bought once. Google Play registration is 25 USD, paid once when the account is created. An app published on both stores therefore carries 99 USD a year permanently, before a single line of code.

The amount is not dramatic. What it implies is. If the Apple membership lapses, the app disappears from the App Store. I often find that renewal attached to the card of someone who left the company months earlier. It is the kind of administrative detail that withdraws a product from sale without anyone having decided to.

The key difference between the two stores is worth noting here: Google's 25 USD buys an account for good, while Apple's 99 USD is a subscription to remaining published. One is a purchase, the other is rent.

The annual cycle that sets your maintenance budget

The real driver of recurring cost is not Apple, it is the Android release calendar. Google enforces a minimum target API level, and it rises every year. Right now new apps and updates must target Android 16 (API level 36), and existing apps must reach Android 15 (API level 35) by 31 August 2026, with an extension available to 1 November 2026 (Google Play Console Help, 2026).

The most important thing about that deadline is that it is not yours. It takes no account of your budget year, your product priorities, or the fact that the app works perfectly. An app delivered in January carries a dated update debt before it has been published once.

In practice this turns maintenance into an annual budget line rather than an unpleasant surprise. It is also the first line cut in a negotiation, which explains a good share of the apps that quietly become unreachable after two or three years. I go into it further in a mobile app needs maintenance like a car.

What actually happens if you do nothing

Both stores penalise inaction, but differently, and the key difference is visibility. Google removes nothing — it simply stops showing you. The wording is explicit: apps below the required level "stop being discoverable to all Google Play users whose devices run Android OS versions newer than your app's target API level" (Google Play Console Help, 2026). Nothing breaks, no message arrives, and installs fall on the newest devices, which are the ones people are buying.

Apple does the opposite, with notice. An app not updated for 3 years and failing to meet a minimum download threshold receives a warning, then has 90 days to submit an update before removal (Apple App Store Improvements, 2026). Both conditions must hold: a rarely-updated but genuinely used app does not fall under this rule.

Google's penalty is the more expensive of the two precisely because it is silent. Nobody tells you a source of installs has closed.

In practice this is what makes the annual update non-negotiable rather than advisable. An app that skips one cycle is not broken and does not generate a single support ticket — it simply stops being offered to anyone who bought a phone that year. By the time the drop shows up in a quarterly report, two cycles have usually passed, and the work needed to catch up is larger than the two updates that were skipped to save money.

The sector cost nobody budgets for at the start

Some sectors add a constraint that is decided during architecture rather than development — and that costs a great deal when it is discovered late. Health is the clearest example in France, because the rule applies to where the data lives rather than to how the app is written. Changing host afterwards therefore means reopening decisions made on day one.

Any organisation hosting personal health data collected during prevention, diagnosis, care or medical-social follow-up must use an HDS-certified host. Certification is issued after a two-stage audit, is valid for 3 years, and carries annual surveillance audits (Agence du Numérique en Santé, 2026; Decree 2018-137 of 26 February 2018).

The real cost is not the hosting price. It is that the list of eligible hosts is short, so the technical choice is constrained from day one. A health app designed without that constraint and brought into compliance afterwards is paid for in rewriting, not in subscription.

What actually decides the budget

In short: account fees are negligible, enforced maintenance is annual and dated, and sector constraints are settled at the beginning. None of those three depends on how many features you build, which is why two apps with an identical starting budget can differ by a factor of two in their third year.

What does depend on it is the surface you have to maintain. Every feature shipped becomes something to retest against every iOS and Android release. That is the strongest argument for a short V1, and it is a budget argument before it is an aesthetic one — the subject of the specification trap.

One figure to place the stakes: across 2,200 apps and 1.3 billion installs, 46.1% of Android installs were uninstalled within 30 days, most on the first day (AppsFlyer, 2025). Paying to maintain features nobody kept the app long enough to see is the most common and least visible expense in this trade.

Frequently asked questions

What is the minimum cost of keeping an app online?

99 USD a year for the App Store, and nothing recurring on Google Play after the initial 25 USD (Apple, 2026; Google Play Console Help, 2026). On top of that, at least one technical update per year to keep pace with Android's target API level. An app with no maintenance budget stays online, but becomes progressively invisible on recent devices.

What happens if I never update the app?

On Google Play it stops appearing for users whose device runs a version of Android newer than your target. On the App Store it may be removed after 3 years without an update if it also fails a minimum download threshold, with 90 days' notice to react (Apple, 2026). Users who already installed it keep it.

Does an App Store rejection cost time?

Apple states that 90% of submissions are reviewed in less than 24 hours (Apple App Review, 2026). A clean submission is usually approved the next day. A rejection sends you back to the end of the queue, and a fix-then-resubmit cycle routinely turns a one-day wait into four or five. The cost of a rejection is a delay, rarely an amount.

An app's budget is therefore not decided at the quote, but in two places that are rarely examined: the calendar the stores impose, and the number of features that will need retesting every year. The first is not negotiable. The second is — as is the decision to ship on one platform or two.