Written by: Nicole Ogonowska, IT Growth Manager, Digital Colliers
Apple just proposed a new commission structure on US linked-out purchases following Epic v. Apple: 15% for standard apps, 10% for partner programs and renewals, and 5% for Small Business Program apps, per 9to5Mac's reporting on the court filing. If any part of your DTC funnel touches iOS, whether that's a native app, a web-view checkout linked from an app, or a subscription managed off-platform, your unit economics have a new variable. And it's not a flat one. It bends based on which tier you sit in, whether the order is new or a renewal, and how you route the user out of the app.
The tiers are not the story, the routing is
Apple's tiers look tidy on paper. In practice, the operational question is which of your orders qualify as "linked-out" and therefore trigger commission at all. A user who taps a button inside your iOS app and lands on your Shopify checkout is in scope. A user who opens Safari on their own and types your URL is not. A user who follows an email link on an iPhone is a gray area that the accounting team will fight about for months.
That matters because DTC margins are already thin. CAC across DTC brands has risen roughly 40% since 2023, Meta CPMs kept climbing through 2024 and 2025, and UK eCommerce only grew about 3% in 2024 versus 2023. Single-digit top-line growth plus a fresh 5 to 15% commission on a slice of orders is the kind of change that quietly pushes borderline SKUs underwater. Profitero's work suggests roughly 30% of SKUs at a typical multi-channel brand already lose money per order once you net out returns and ad spend. A new iOS commission line item makes that worse before it makes it better.
What you actually need to compute
To know your exposure per order, you need to answer four questions at the moment of purchase:
- Did this session originate from an iOS app you own, or from an iOS app surface Apple considers in-scope?
- Which commission tier does your developer account sit in (5, 10, or 15%)?
- Is this order a new purchase, a subscription renewal, or a partner-program transaction?
- What's the commissionable base (gross, net of tax, net of shipping, net of refunds within the window)?
Most DTC stacks answer none of these cleanly today. The attribution layer knows the ad click. The checkout knows the order. Nobody stitches "this order came from an iOS app link-out" onto the order record in a way finance can audit six months later.
The data model, concretely
The shape of the fix is boring and that's the point. On every order, stamp three fields:
ios_linkout_flag(boolean, resolved at session start from the referrer or a signed token Apple's SDK provides)apple_commission_tier(enum: 5, 10, 15, or null)apple_commission_basis(the dollar amount Apple's rule applies to, computed at order close and frozen on refund events)
Then a nightly job reconciles those against Apple's reporting once it exists, the same way you reconcile Stripe payouts today. Refunds inside the commission window need to write back and reduce the accrued commission. Renewals need their own flag because the 10% tier applies differently. If your subscription product bills monthly on iOS-originated signups, month 13 is a different rate than month 1, and your LTV model needs to know that.
The hard deadline
Apple's proposal is tied to the Epic v. Apple remand and is live in the US now, with the court giving Apple until mid-January 2026 to finalize compliance mechanics. That's the window. Operators who wait until Q2 to instrument this will be back-filling commission accruals from memory and CSV exports. Operators who ship the three fields above before the January cutoff will close their January books with iOS commission as a clean line item, not a mystery variance.
The teams I keep seeing get this right treat it as a finance data problem, not an app problem. The app change is a few hours. The data model, the reconciliation job, and the LTV recompute are the actual work. Start there, and the 15/10/5 becomes a number you can price against instead of a surprise on the P&L.