Apple’s New App Store Commission Proposal: What 15% Actually Costs a Subscription App

Apple filed a new commission proposal with the US District Court for the Northern District of California on 2026-08-13: 15% on external link-out purchases for standard apps, 5% for apps in the Small Business Program, and 10% for apps in Apple’s Video, News, and Mini Apps Partner Programs plus subscription renewals. Every outlet that covered the filing reported the tier structure. None of them ran the numbers for what a subscription app actually pays under it, at more than one revenue size, or what happens to the store’s books once a second payment rail exists alongside Apple’s own. This post does both: a line-item cost model across three revenue tiers, and the reconciliation consequence that shows up in NetSuite the moment checkout splits across two processors.

On this page

• What Apple actually proposed, and why now
• Today’s rate is zero — and that’s not permanent
• IAP vs. external link-out: the worked numbers
• Which tier applies to your app
• If you’re already routing externally, this is your real number
• The cost the coverage misses: two revenue streams, one GL
• Before you route checkout externally

What Apple actually proposed, and why now

The filing is the latest turn in Epic Games, Inc. v. Apple Inc., the case that has run since 2020. A Ninth Circuit panel upheld the district court’s April 2025 civil-contempt finding against Apple but ruled that the original order — a flat ban on any commission for external link-out purchases — was overbroad. It remanded the question with a new standard: Apple may charge only for costs it “genuinely and reasonably” incurs in coordinating those purchases, not a market-rate commission. That “necessary costs” language is doing the real work here — Apple’s own filing concedes that a strict reading of the standard would put the fee at 0%, and proposes 15/10/5% anyway on the argument that platform value justifies more than bare cost recovery.

Epic Games rejected the proposal outright and has said it will file its own opposition. Judge Gonzalez Rogers now has to weigh two positions that don’t overlap — Apple’s cost-plus-value number against Epic’s cost-only number — with no ruling yet on the docket. The tiers themselves are not new territory: 15% is exactly half of the standard 30% in-app-purchase commission, and 5% is a third of the 15% Small Business Program rate that’s applied to IAP since 2021. Apple is proposing to keep the same ratio, just on a different revenue rail.

This filing covers US transactions only — the EU already runs its own, separate fee structure under the DMA, not addressed here.

Today’s rate is zero — and that’s not permanent

The number that matters for a decision made this week is not 15%. It’s 0% — what Apple is actually charging on external link-out purchases right now, and has been since the April 2025 contempt ruling took the fee to zero while the case worked through appeal.

That’s easy to misread from older coverage: Apple’s pre-injunction external-purchase commission ran as high as 27% once its own processing fee was layered on top of the base rate — a number still cited in some writeups as if it were current. It isn’t. The 27% figure describes a structure Apple has not been allowed to charge since April 2025; 0% describes today; 15/10/5% describes a proposal still awaiting a ruling.

That distinction is the actual planning problem. Any app that implemented Apple’s External Purchase Link entitlement after the original 2024 ruling is paying only its payment processor today, not Apple. If the court adopts anything close to Apple’s proposed tiers, that changes on the date the order takes effect — not gradually, and not with much notice built into the process. The rest of this post treats 15/10/5% as the rates to plan against, because they’re the only concrete numbers on the table, while treating “not yet in force” as a fact that should shape when you act on them, not whether you model them.

IAP vs. external link-out: the worked numbers

The comparison below is a modeled example, not a real customer: a subscription iOS app at a flat $10/month price, no trials or discounts skewing the mix, three revenue sizes, and a standard US card-processing rate for the external-checkout leg. Swap in your own price point and processor and the ratios hold even if the dollar figures don’t.

Two commission numbers apply to every tier: staying inside Apple’s in-app purchase system costs 30% standard or 15% under the Small Business Program (eligibility is based on the prior calendar year’s proceeds — revenue net of Apple’s own commission — staying under $1,000,000, not gross sales). Routing externally at Apple’s proposed rates costs 15%, 10%, or 5% depending on the same eligibility test, plus whatever the external payment processor charges on top. This model uses Stripe’s published US standard rate, 2.9% + $0.30 per transaction, as the processor assumption — Braintree, Adyen, and regional processors price differently, and that difference moves the external-side total in either direction.

Verdict: at every modeled tier, routing external still beats staying in IAP on commission alone once the proposed rates are in force — roughly a 27–30% reduction in total platform-plus-processing cost. That gap is the ceiling, not the answer: it assumes the processor fee modeled here, ignores the engineering and support cost of running a second checkout, and says nothing yet about what happens to the money once it lands. The next two sections cover both.

Which tier applies to your app

Two questions decide the rate, and they’re not independent — Small Business Program status is checked first, because Apple’s proposal applies it as a flat override regardless of app category.

If you’re already routing externally, this is your real number

An app that already implemented Apple’s External Purchase Link entitlement is not comparing 15% to 30% today — it’s comparing 15% to 0%. Every dollar of the current savings from routing external comes from the fact that Apple isn’t charging anything on that rail right now, not from the processor being cheap. The number worth tracking is the marginal cost of the proposal actually being adopted: at the $50,000/month tier modeled above, that’s the jump from $2,950/month (processor fee only, today) to $5,450/month (processor fee plus Apple’s proposed 5%) — an added $2,500 a month that shows up the day an order takes effect, on top of whatever the app is already paying.

At the $200,000/month tier, the same jump is $11,800 to $41,800 — an added $30,000 a month. Budget for that jump before it happens; the court’s calendar, not the news cycle, decides when it lands.

The cost the coverage misses: two revenue streams, one GL

Every article on this filing stops at the commission percentage. None of them cover what happens to the money once it’s collected — the part that matters for a store already syncing Woo or Shopify orders into NetSuite. Staying entirely inside Apple’s in-app purchase system means one consolidated payout report, one settlement schedule, one line to reconcile. Splitting checkout — even partially, even just for renewals — means two: Apple’s IAP payout for whatever stays inside the app, and a separate settlement from whatever processor handles the external link-out, arriving on its own schedule with its own fee format.

NetSuite’s own documentation is specific about where its automation stops. Automated Settlement Reconciliation is scoped to payments processed through NetSuite Pay itself — its own payment link, SuiteCommerce checkout, the Customer Center. Neither Apple’s IAP payout nor a third-party checkout processor is a NetSuite Pay transaction, so neither qualifies.

Both fall to what NetSuite calls native bank reconciliation instead, and Oracle’s own support documentation states the gap directly: “Native NetSuite Bank Reconciliation does not take into account transaction fee reconciliation,” while NetSuite Pay’s own Settlement and Fee Reconciliation “automatically reconciles transaction fees” — for its own rail only. In practice that means the commission line inside Apple’s payout and the processor’s fee line on the external settlement both have to be pulled out and matched by hand, or by a rule you build yourself, rather than by the automation NetSuite reserves for its own checkout.

None of that is a reason to avoid routing externally — the commission math above still favors it at every modeled tier. It’s a reason to price the operational cost alongside the commission cost before deciding. A store that already has clean reconciliation on a single payment rail is adding a second rail with its own settlement timing, its own tax treatment, and its own manual-match burden — real ongoing work, not a one-time setup task, and it’s the exact gap none of the coverage of this filing has mentioned.

Before you route checkout externally

• Confirm Small Business Program eligibility against last calendar year’s actual proceeds, not this year’s target — the test is retroactive and re-qualifies annually.
• Check that qualifying purchases land inside Apple’s existing 7-day link-tap attribution window — a purchase outside it reverts to standard in-app commission, not the external rate.
• Price your actual processor’s rate card instead of assuming 2.9% + $0.30 — Braintree, Adyen, and region-specific processors quote differently, and the gap changes the verdict at the margin.
• Map which general ledger account each revenue stream posts to before the first external transaction settles, not after the first month-end close surfaces the gap.
• Assign sales-tax and VAT remittance ownership explicitly — Apple’s marketplace-facilitator role covers in-app purchases; a purchase routed around it is not automatically covered the same way.
• Build, or manually run, a second fee-reconciliation pass. NetSuite’s automated settlement-and-fee matching is scoped to its own payment rail; a second income stream lands through native bank reconciliation, which does not auto-match fees.
• Model the scenario where the court sets a different number than Apple proposed. Epic is asking for 0%; nothing here is final until Judge Gonzalez Rogers rules.

Every item on that list is cheaper to work through before the first external transaction closes than after — the reconciliation gap in particular tends to surface at month-end close, well after the checkout decision was already shipped.

If routing checkout externally is on the table, the reconciliation question above is worth answering before the commission decision, not after — an e-commerce sync audit is built exactly to map every revenue stream feeding your general ledger — App Store payouts, external processors, storefront orders — and flag where they don’t reconcile cleanly, before a quarter close does it for you. The same fee-math discipline used here applies just as directly to marketplace fee comparisons and to pricing custom integration work — both covered elsewhere in the NetSuite and WooCommerce integration guide library. A payment-processor concentration event is already in the news this week too — see the Stripe-PayPal acquisition talks and what they mean for merchants running either processor.

Frequently asked questions

Related guides

• How to Choose a NetSuite Integration Partner (Without Burning $40k)
• Composable Commerce and NetSuite: Adding Services Does Not Add Concurrency
• Vibe Coding Is Real: The Subscription Is a Balance, Not a Seat