
Published date:
September 11, 2026
Apple's 2026 EU terms cut App Store fees to as low as 5% and open up web billing, where a $10 subscription keeps roughly 38% more margin. Here is how subscription apps should split revenue between store and web, and how to migrate without losing the margin.
For a decade the mobile subscription business ran on a rule nobody voted for: if you wanted to reach an iPhone or an Android user, you billed through Apple or Google and handed over up to 30% of every payment. That arrangement is now being taken apart by regulators, courts and the developers themselves, which puts subscription payment processing at the centre of the growth conversation.
Apple's August 2026 changes in the EU are the clearest signal yet that the closed-billing model is ending, and they force a question every subscription business should already be modelling: how much of your revenue belongs on the web, and how much still belongs in the store.
This guide covers what changed, why businesses are moving revenue to the web, how to decide where each subscriber cohort should be billed, and how to make the move without giving back the margin you set out to capture.
On 18 August 2026 Apple announced a single set of EU business terms for developers, effective 1 October 2026, replacing the fragmented model it had operated since the Digital Markets Act took hold. Two structural shifts matter for anyone selling subscriptions.
First, the Core Technology Fee is gone. The per-install charge Apple introduced in 2024 as its DMA-compliance mechanism has been replaced by a Core Technology Commission of a flat 5% on digital transactions in apps distributed outside the App Store, which swaps a fixed per-install cost that punished high-volume free-to-paid funnels for a revenue-linked charge.
Second, Apple In-App Purchase and alternative payment options can now co-exist in the same EU app, which was not previously permitted and is the change that makes a genuine mixed model possible.
The new commission ladder looks like this:
The reduced rate is not applied by default. A transaction qualifies for it when at least one of these conditions is met:
That ladder is the product of a multi-year fight that Apple lost in stages. The Digital Markets Act, Regulation (EU) 2022/1925, became applicable in March 2024 and required gatekeepers to allow steering to outside payment, and enforcement moved fast from there:
The pattern is consistent on both sides of the Atlantic, and Britain's CMA reached the same conclusion when it designated Apple and Google with Strategic Market Status in 2025: the anti-steering rules that once made app-store commissions unavoidable are being struck down, and Apple's own August 2026 statement says the new terms resolve its disagreements with the Commission.
Google is on the same path, having settled an antitrust claim with Match Group in 2023, and from June 2026 it is restructuring Play fees so that routing payments to an external checkout cuts its take to around 10%, down from the 15% to 30% it charged through Google Play Billing.
App stores take between 15% and 30% of every subscription payment, while web card processing costs a fraction of that, typically around 2.9% plus a small fixed fee. On a $10 subscription, Apple's 30% standard rate leaves the developer $7.00, whereas a 3% web fee leaves about $9.70, which is roughly 38% more retained margin on the same sale. At the 15% tier the gap narrows, but web still keeps meaningfully more of every dollar.
The prize is large: Appfigures estimates that Apple alone collected more than $10.1bn in US App Store commissions in 2024 and over $27.39bn globally, and the subscription economy was valued at around $492bn in 2024 with double-digit annual growth projected for the decade ahead.
The businesses acting on that arithmetic are among the largest in the market:
Scale is the pattern behind these moves. RevenueCat's State of Subscription Apps 2026, built on more than 115,000 apps, found that 41% of top-tier apps generate web revenue against just 1.3% of the smallest hobby apps.
Web does not win on every metric, and being precise about that keeps the case credible. The benchmarks split cleanly:
The strongest control advantage is involuntary churn, meaning the revenue lost when a card fails rather than a customer choosing to leave:
The old shorthand that web escapes the 30% cut no longer holds under the new EU rules, so a serious strategy starts from what each channel actually does well.
The saving depends heavily on how long a subscriber has been paying, because that sets the in-app rate you are comparing web against. A year-one subscriber is billed at the 26% standard rate, so moving that cohort to web captures close to the full gap down to roughly 3% processing. A second-year subscriber already sits at the 15% reduced rate, so a 15% link-out saves little over staying in-app. The margin case is therefore strongest on year-one subscribers and high-ARPU annual plans, and thin on long-held subscribers.
One caution before going all-in on web: if annual ARPU is below roughly $60 and you cannot identify a distinct web audience, vendor data suggests a forced web-to-app funnel can lose more in conversion than it saves in fees.
Combining the two channels is now possible in the EU, where a developer can offer Apple In-App Purchase alongside alternative processing and web link-out in the same app. The catch is that payment options must be selected and held for 12 months, so the blend is a planning decision rather than a monthly dial.
A blended model also has to be executed across in-app purchase, card, Apple Pay, Google Pay and direct debit, in several currencies, on cycles you may want to change, which is where a subscription payment gateway earns its place. COLIBRIX ONE recurring payments runs fixed, usage-based and hybrid billing through a single API and lets you adjust cycles and amounts on demand, while accepting Visa, Mastercard, Apple Pay, Google Pay, SEPA and ACH across the US, UK and Europe, so the mix you commit to for the year is one you can actually operate.
Start with the cohort math before touching the migration: split subscribers by tenure and ARPU, because the upside concentrates on year-one and high-ARPU subscribers, whereas a cohort already at the 15% reduced rate may not clear the cost of running your own billing. Sequence the build from there, standing up the checkout and recovery stack first, the funnel second, and measurement from day one, tracking retained margin per cohort, the conversion delta against your store baseline, retention at month one and month six, and involuntary-churn recovery.
A recurring payment is a charge a customer authorises once and a business then collects automatically on a fixed or usage-based cycle, which is the billing model behind almost every subscription. Moving it to the web means you, rather than the app store, own that mandate and the customer data around it.
The headline card processing fee is around 2.9% plus a small fixed charge, but the all-in cost of owning billing is higher, since you also absorb fraud, chargebacks, refunds, PCI compliance, sales-tax and VAT registration and filing, and billing-support headcount. One analysis puts the true blended cost as high as 7.8% for a global SaaS running a full stack, which still sits well below 15% to 30% and is the figure your business case should use.
You need to charge in several currencies, accept the methods your subscribers actually use, and match every payment back to a subscriber without manual work, which is why most teams run this through a subscription payment gateway rather than build it in-house. COLIBRIX ONE recurring payments executes fixed, usage-based and hybrid billing through a single API and accepts Visa, Mastercard, Apple Pay, Google Pay, SEPA and ACH across the US, UK and Europe, with real-time transaction tracking for reconciliation and onboarding completed in around 48 hours.
Own the recovery levers the platforms otherwise run on their defaults: your own dunning schedule and retry timing, pre-dunning notices, and card-updater services such as Visa Account Updater and Mastercard Automatic Billing Updater. This matters because involuntary churn, meaning the revenue lost when a card fails rather than a customer choosing to leave, runs at 24% of overall churn for consumer subscriptions per Stripe's own benchmarks, and under app-store billing those levers sit with the platform.
In the EU you now can, since a developer may offer Apple In-App Purchase alongside alternative processing and web link-out in one app. Two rules shape the design: payment options must be selected and held for 12 months, so the blend is a planning decision rather than a monthly dial, and the funnel must respect Apple's presentation requirements and its under-18 restrictions, including parental gates and no external links in Kids category apps.
There is no single answer, but the practical test is whether the system runs any billing model through one API, supports the payment methods and currencies you sell in, and gives you the recovery and reporting levers that protect revenue once you own billing. Weigh it against the cohort math above, because the right setup for a high-ARPU annual plan is rarely the right one for a low-ticket monthly plan.

Not all IBANs work the same way: a shared IBAN forces manual reconciliation and reference-code matching...

As subscription businesses scale, simple billing processes often break under the weight of payment failures and technical debt...

Total global B2B payment volume is projected to exceed $280 trillion by 2034, reflecting the scale and importance of modern payment...