On October 1 Apple hands EU developers a choice it has never offered before, and then takes the undo button away for a year.
Almost all the coverage of Apple’s unified EU business terms has been a rate card. Twenty-six percent instead of thirty. A five percent Core Technology Commission instead of a fifty-cent per-install fee. Those numbers are real, and we have written about why the fee-instrument switch matters more than the headline rate. But a rate card is something you read once. The part of these terms still shaping your P&L next August is a single sentence about duration, and it has been covered almost nowhere.
This piece is about that sentence: what Apple binds you to, the four things Apple has not published about it, and how a user acquisition team should make the election — because the configuration you pick sets net revenue per EU install for twelve months, and net revenue per install is the number every bid ceiling is built on.
01What Apple actually requires
Two Apple pages define the obligation. The fuller one, on payment options in the EU, reads:
“To help ensure a consistent user experience, once you select a payment option or combination — Apple In-App Purchase, alternative payment processing within the app, and/or out-of-app offers with actionable links — you must maintain that choice across all EU storefronts for 12 months.”
The EU hub page says the same thing while enumerating the menu slightly differently: developers “will select their payment options — Apple In-App Purchase, alternative payment processing in their app, linking to the web, or a combination — and must maintain those options for 12 months.”
Three things follow. Coexistence is genuinely new: “Apps distributed in EU storefronts can now offer alternative payment methods and offers alongside Apple In-App Purchase.” Under the previous regime you could not run Apple’s rails and someone else’s side by side. The binding scope is all EU storefronts, so there is no per-country election. And the clock does not start on a common date: an account becomes subject to the terms “starting October 1, 2026, or the date they agree, whichever is later.” Two competitors in one category can be twelve months apart on the same decision.
One structural note: the support pages are not the contract. Apple says these terms “will be superseded by Attachment 14 of the Apple Developer Program License Agreement” — and that attachment, where renewal and exit mechanics would live if they live anywhere, is not publicly readable.
02The rate card, and the fee that got renamed
Three payment paths, each with a standard and a reduced rate, all effective October 1, 2026.
| Payment path | Standard | Reduced |
|---|
| Apple In-App Purchase | 26% | 15% |
| Alternative processing in the app | 20% | 10% |
| Out-of-app offers (7-day window) | 15% | 10% |
| Distribution outside the App Store | 5% | 5% |
The reduced column is one shared eligibility clause, worded identically in all three rate tables, not four separate rules: participants in the App Store Small Business Program, the Mini Apps Partner Program or the Video Partner Program, plus auto-renewable subscriptions after their first year.
Now the trap that produces a published error if you are careless with it. Apple states that the new terms “eliminate the Initial Acquisition Fee and Store Services Fee” — the old 2% and the old tiered 5%/13%. True. Apple simultaneously introduces a store services commission of 15% on out-of-app offers. Also true, same rate card. Write “Apple eliminated the store services fee” and stop, and every developer reading you concludes link-outs are free. They are not.
That commission has a boundary user acquisition people will recognise instantly, because it is an attribution window: “Only sales made within 7 days of the link tap are subject to this commission.” Seven days from tap. Whatever else you think of the charge, its shape is a click-attribution model, and its behaviour under a long consideration cycle is worth modelling before you choose it.
→ THE NOUN IS THE ARGUMENT
“Steering fee” is the coalition’s word. Apple says “store services commission.”
Apple’s rate-table row for the 15% charge reads simply “Out-of-app offers,” and Apple does not use the term steering fee anywhere. The disagreement over the noun is the whole Article 5(4) dispute in miniature: a commission is payment for a service rendered, whereas a fee on steering is precisely the thing the DMA says must be free of charge. Which word you use is a legal position, not a style choice.
03Four things Apple has not published
We read both pages that define the lock, end to end, on September 14. Four questions any competent finance team will ask are not answered on either.
Is the election per app or per account? Apple never says. Both pages use the second person — “once you select” — and neither attaches the choice to a bundle ID or a developer account. The only scope stated is “across all EU storefronts,” which rules out a per-country split and settles nothing else. Some trade coverage asserts per-app; we could not find that claim sourced to Apple wording, and we will not repeat it. If you run a portfolio, this is the question worth getting in writing before you agree.
What happens at month twelve? Verified absent. No renewal mechanism, no re-election window, no notification, no stated process for choosing differently when the period ends.
How do you change the configuration inside the period? No process is published. Not a restrictive process — no process at all.
Is there an early exit? No early-exit, hardship or amendment provision appears on either page.
None of this makes the terms unusable. It does mean the honest way to plan is to treat the twelve months as genuinely irreversible, which is a materially different posture from “we can revisit in Q2” — and a familiar shape for anyone who has watched Apple ship a consequential change with no changelog attached, which we keep a running file of.
04The install base you are locking into
This is the constraint nobody has written about, and it is the strongest argument here.
Apple states that the entitlement enabling alternative payment processing “may only be used in apps on EU storefronts on devices running a minimum of iOS 26.2, iPadOS 26.2, macOS 26.6, tvOS 26.6, visionOS 26.6, and watchOS 26.6.” For reporting transactions back through the External Purchase Server API, the floor is higher still: iOS 26.4 and later.
Note what that is not. Alternative distribution — marketplaces — has a floor of iOS 17.4 or iPadOS 18, by now nearly everyone. Alternative payments inside an App Store app has a floor of an OS released a few months ago. Coverage uses the two interchangeably. They are not.
iOS 26.2+
The floor for the alternative-payments entitlement — so the configuration you freeze for twelve months has an addressable base on day one that is a fraction of the EU fleet, and that grows underneath you all year while your choice does not move.
Work the consequence through. Elect alternative processing alone and you serve a 20% path to the supported slice with no Apple IAP fallback for everyone else. Elect a combination and you carry two integrations from day one, getting the cheaper rate on whatever share of revenue travels it. That is why coexistence was worth granting, and also why the lock bites: the right mix at low device penetration is not the right mix a year later, and you choose once.
A second, smaller version of the same problem sits in the child-safety rules. Apps offering alternative payments in the EU must put purchase flows behind a parental gate for users under 13, and for 13- to 17-year-olds both in-app processing and out-of-app offers must sit behind a gate; Kids category apps cannot link out at all. Apple’s implementation note: call canMakePayments in StoreKit, and “in a future software update, Apple will release new APIs to better support these requirements.” The obligation binds on October 1. The APIs to satisfy it do not exist yet.
05What the lock does to your acquisition maths
Here is the connection Apple does not make and no vendor has made either. Apple’s four pages defining the October 1 regime contain no mention of Apple Ads, advertising, ROAS or acquisition economics, and the most thorough third-party explainer we could find offers no guidance on how to make the election. Nothing about the transition changes Apple Ads billing or reporting on October 1 — which is worth knowing precisely because it means nothing will prompt you to revisit your bids.
The arithmetic is unavoidable, though. Your maximum affordable cost per tap is net revenue per download, multiplied by your tap-to-download conversion rate, divided by your payback multiple. The payment path sets the first term.
Flagged as derived illustration, not measured data. Take €1.00 of gross digital revenue attributable to an EU install. On the standard IAP path at 26%, net is €0.74. On in-app alternative processing at 20%, net is €0.80 before your own costs — call it €0.77 after a typical 2–3% PSP charge. On a link-out at 15%, net is €0.85 before processing, roughly €0.82 after. At a 60% conversion rate and a 1.0× payback multiple, those give tap ceilings of about €0.44, €0.46 and €0.49 — standard to link-out is roughly an 11% lift in what you can afford to bid.
Three caveats matter more than the numbers. The lift applies only to the share of revenue that actually travels the cheaper path, bounded on day one by that iOS 26.2 floor. It is reduced by your own processing, tax and fraud costs, which Apple was previously absorbing. And it is zero for anyone already at 15% — a Small Business Program participant sees no IAP-path change at all.
One more acquisition-specific wrinkle, and it is a good one. Apple’s rule: “Your app’s App Store product page may not include information about purchasing with an alternative payment option.” The product page is the acquisition surface — what a declared-intent searcher lands on after tapping your ad, and the reason App Store search buys the highest-value users on iOS. So the surface where you convert intent is the one surface on which you may not mention the cheaper path you just locked into. Same rendering-layer logic we traced when Apple banned pricing from creative assets: the offer is not forbidden, its location is dictated.
→ FREE 10-POINT AUDIT
Get your EU baseline on record before October 1
We split your Apple Ads reporting by EU storefront, compute your current net revenue per install and the tap ceiling it implies, and show you what each payment path does to that ceiling — so you make the twelve-month election with the acquisition maths attached. One hour. Senior strategist. No pitch.
Book my free audit →
06Locking into terms the regulator has not cleared
Apple describes the new terms as the product of “close collaboration with the European Commission.” The Commission has made no reciprocal statement. That asymmetry is the whole story here.
The Commission’s own annual DMA implementation report, COM(2026) 247 final, dated May 21, 2026, records that it “issued a preliminary finding that Apple had failed to comply with Article 6(4) DMA regarding the contractual terms,” and notes in a footnote that “the case is still ongoing as of March 2026.” Article 6(4) governs alternative distribution. Separately, on April 23, 2025 the Commission fined Apple €500 million for breaching the Article 5(4) anti-steering obligation; Apple filed for annulment on July 7, 2025 and that appeal is undecided.
Into that, on September 8, 2026, a coalition of European trade bodies sent an open letter to the Commission. Reported signatory counts differ — MacObserver and MobileGamer say eighteen organisations, Mobile Marketing Reads says seventeen, and no outlet published a complete list — so take the number as approximate. Named signatories include the Coalition for App Fairness, the European Games Developer Federation, the European Publishers Council, France Digitale, News Media Europe, Euroconsumers and Uptodown. The letter argues that Apple’s “revised framework does not resolve Apple’s non-compliance with the Digital Markets Act (DMA) and continues to undermine meaningful competition,” and that “excessive fees and restrictive conditions risk preventing alternative distribution channels from developing into effective competitors.” It asks the Commission to consult developers before accepting the October 1 terms, to clarify which proceedings remain open, and to conclude them. Apple has not publicly responded.
We could not locate the primary text of that letter on any signatory’s own site, so every quotation above is second-hand from trade coverage. Which is itself the point: the letter was covered by mobile-industry trade press and by nobody else — no wire service, no Brussels policy outlet. A regulatory challenge to terms binding thousands of developers in seventeen days is, so far, invisible outside the developer trade.
The practical reading is not that the terms will be struck down. It is that you are being asked to make a twelve-month irrevocable election inside a framework whose predecessor the regulator has formally declined to bless, with no published mechanism for responding if it changes mid-period.
07Nobody knows what anyone did last time
If you wanted to reason from precedent — how many developers took alternative distribution when it first became available, how much volume moved — you cannot, because nobody has measured it. Apple publishes no marketplace count, no developer count and no Core Technology Fee revenue figure. More strikingly, neither does the regulator: that same annual DMA implementation report, the document whose job is to assess whether the regime is working, contains no quantified data on marketplaces operating, developers on alternative terms, or volume moved off the App Store. Twenty-eight months after alternative distribution became legally available, the enforcing authority published an annual report that does not measure whether anyone used it.
The closest thing to a census is journalism. TechCrunch counted seven EU alternative marketplaces in February 2026 — AltStore PAL, Setapp Mobile, Epic Games Store, Aptoide, Mobivention, Skich and Onside — and reported that Setapp Mobile had closed on February 16. Roughly six or seven ever launched; at least one is already gone. No developer counts, no download figures, no share shifted.
It is the same evidentiary vacuum we keep hitting on this channel: no post-multi-slot Apple Ads benchmark exists either, and no published Apple Ads incrementality result exists anywhere. You are not being asked to decide with imperfect data. You are being asked to decide with none.
08How to actually make the election
Six steps, ordered so the reversible work happens first.
One: split EU reporting now. You need a clean pre-transition baseline of revenue, installs and cost per install by EU storefront, and after October 1 it will not exist to take.
Two: compute the share of EU revenue that can actually travel each path. Not your total EU revenue — the share on iOS 26.2 or later, and the share from payers rather than installers. That is the multiplier on every rate saving, and for most apps it will be smaller than the rate difference suggests.
Three: price your own stack honestly. Processing, fraud, chargebacks, EU VAT, refunds, support for payment failures. Apple absorbed all of it at 26%. If your fully loaded cost is 4%, the 20% path nets 24% and the gap closes to almost nothing.
Four: ask Apple, in writing, whether the election binds per app or per account. If you run a portfolio and it binds account-wide, this is a different decision.
Five: default to the combination unless you have a reason not to. It is the only configuration that does not require you to predict device adoption twelve months out. Apple polices it: alternative processing offered alongside out-of-app offers must be “viewable and selectable on the same screen” in a manner that does not discourage or obfuscate its use.
Six: do not move bids on October 1. Net revenue per install changes on the day; your measured conversion data does not, and the honest bid-ceiling calculation needs a payer-mix input you will not have for a month. Re-derive ceilings in November with real data. The same restraint applied after the September attribution change: when the measurement floor moves, let it settle before you move on top of it.
The gaps we are naming
Apple has not stated whether the twelve-month election binds per app or per developer account, what happens when the period ends, how to change the configuration inside it, or whether any early exit exists; the binding text sits in Attachment 14, which is not publicly readable. The APIs Apple says are needed for its own child-safety conditions arrive “in a future software update,” after the obligation binds. The primary text of the September 8 letter could not be located on any signatory’s site, and the signatory count is reported inconsistently. No party — not Apple, not the Commission in its own annual report, not any analytics vendor — has published adoption data for the previous EU terms, so there is no precedent to reason from. The bid-ceiling figures above are derived illustrations on Apple’s published rates, not measured outcomes, and assume a payer mix and conversion rate you should replace with your own. And no EU-specific Apple Ads benchmark set exists, so even after October 1 the effect on what an iOS install actually costs in Europe will be measured by you, in your own account, or not at all.