Web shops change the maths of paid UA
A web shop nets a studio far more per dollar than an in-app purchase. That margin gap should reshape payback windows and bidding, carefully.
When a studio keeps ninety cents of a dollar instead of seventy, every number underneath a UA budget needs rebuilding, not adjusting.
Correction (2026-09-20): An earlier version of this article relied on a misstated Apple EU fee chronology (including via cross-reference to our CTC web-shop math piece). Under Apple’s June 26, 2025 update, Apple planned to move from the Core Technology Fee (CTF) to a Core Technology Commission (CTC) in a single EU business model. On August 18, 2026, Apple announced that the per-install CTF is replaced by a 5% CTC on digital transactions in apps distributed outside the App Store, with the new terms effective October 1, 2026, and that the Initial Acquisition Fee and Store Services Fee are eliminated. See Apple Developer News (June 26, 2025; August 18, 2026).
The margin shift, worked through
A purchase inside the app store, on either platform, typically nets a developer somewhere in the region of 70 to 85 percent of the transaction after commission, depending on program tier and jurisdiction. Route the same purchase through a studio's own web shop and, once payment processing and any platform-mandated fee on the referred or out-of-store transaction come out of it, the take commonly sits closer to 90 percent or more. Our companion piece on Apple's Core Technology Commission and web-shop math (corrected 2026-09-20) sets out the EU chronology: CTC under the August 18, 2026 announcement is a 5% commission on digital transactions in apps distributed outside the App Store, effective October 1, 2026 — not a blanket 5% on every digital-goods dollar under every distribution path, and not a finished January 1 swap. The direction of the change is still the same wherever a studio operates a credible web shop. Web shop revenue is often worth more per dollar of gross spend than the equivalent in-app purchase, but the EU fee line items have to be modelled against the verified path and effective date.
That gap of ten to fifteen percentage points doesn't sound dramatic until you run it through a payback model. A campaign priced to break even at, say, a 180-day payback window on in-app revenue can often sustain a materially shorter window on revenue routed to a web shop, or a materially higher acceptable CPI, because the same player spend returns more net cash to the studio. For a studio running meaningful volume, that difference compounds into either more aggressive bidding capacity or a faster route to profitability on the same cohort — after CTC (where it applies), processor fees, and conversion friction are included.
Which cohorts actually belong on the web
Not every player is worth routing to a web shop, and treating the funnel as a blanket redirect usually backfires. There is real friction in leaving the app, logging in somewhere else and completing a purchase through a separate checkout. The users worth that friction are disproportionately the high-intent, high-spend segment: players who have already converted at least once in-app, where a modest discount or bonus currency gives them a reason to move future purchases to the cheaper channel. Push a first-time, low-intent user into a web checkout before they have any trust in the game and you usually lose more to abandoned purchases than you save in platform fees.
The pattern most studios have settled into aims web shop promotion at returning payers and known high spenders; in-app messaging or a push notification routes them to the web shop for subsequent purchases. First purchases and casual spenders stay inside the standard in-app flow, where friction is lowest.
Rebidding, carefully
The temptation once a studio sees the margin improvement is to raise bids across the board, on the assumption that a meaningful share of future revenue will move to the cheaper channel. That's a risky thing to price into a live campaign. Web shop adoption rates vary enormously with genre and region, and with how aggressively a studio markets the option, so a bid raised against an assumed future margin that never materialises erodes payback rather than protecting it.
The safer approach models two payback numbers for the same cohort: one assuming all revenue stays in-app, the other assuming the studio's own measured web shop adoption rate for comparable past cohorts. Bidding moves toward the blended figure only once a studio has enough historical adoption data on its own player base to trust the assumption, not on the theoretical maximum a web shop could reach.
The platform risk that does not go away
None of this changes the underlying platform risk. Both major platforms have adjusted the terms attached to external purchase links more than once since those links became available, and a studio building bidding models around a web shop margin advantage is implicitly betting that the current fee structure holds for the length of its payback window. Apple's August 18, 2026 EU package (effective October 1, 2026) is one reminder that fee labels and effective dates can move after a planning model is already live. So platform policy monitoring belongs on the recurring line rather than in a one-time compliance check, for any studio with a meaningful share of revenue now riding on web shop economics. The maths above rests on terms a platform can revise with notice measured in weeks rather than years.
The studios handling this well keep the web shop margin advantage out of their headline bidding targets and build it into a separate, clearly labelled adjustment layer instead, one they can dial back quickly if a platform tightens terms or if measured adoption falls short of what a cohort's model assumed. That separation matters more than the size of the advantage itself. A studio that can't isolate how much of its bidding capacity depends on web shop economics won't notice the moment those economics change, until payback across the whole portfolio has already slipped.
