Apple Still Enforces the Rules
Apple temporarily removed Cal AI from the App Store last week for multiple violations related to payment handling and deceptive billing practices. The calorie-counting app — which MyFitnessPal acquired in March at $50 million in ARR — had attempted to bypass wiki:in-app-purchase requirements by implementing an embedded Stripe payment flow that removed Apple's IAP as a checkout option.
This violated Guideline 3.1.1, which requires that IAP be offered alongside any external payment link — even following the Epic v. Apple ruling that permits U.S. developers to link out to web payments. The exception is "reader apps" (those providing subscription-based access to digital content like books, audio, or video). Cal AI does not qualify.
Apple also cited manipulative tactics: the paywall displayed weekly pricing more prominently than the actual billing amount, obscured subscription auto-renewal behind a free trial toggle, and prompted users who declined the first subscription offer with a second, different purchase flow. User reviews accused the app of being a scam.
Cal AI has since addressed the violations and returned to the App Store. The message is clear: even after Epic loosened some restrictions, Apple actively polices how developers implement external payments. For high-growth apps looking to skirt platform rules, the risk remains enforcement at scale.
The App-to-Web Promise vs. Reality
The surface-level economics of app-to-web checkout are seductive: avoid the 15–30% platform commission, keep more revenue per paying user, and regain control over the billing relationship. The U.S. Epic ruling opened the door. Payment infrastructure providers now support the technical path. Yet the full-funnel reality is messier than the margin math suggests.
Conversion typically drops when users are sent out of the app, into a browser, through another authentication context, and back via deep link or redemption flow. Every additional step introduces friction, context switching, and abandonment risk. Even well-designed web checkout flows are more fragile than native wiki:in-app-purchase paths. If enough users fail to complete checkout, the margin recovered on successful conversions evaporates.
One publicly documented experiment saw 6% fewer paying customers when web checkout was added. Interestingly, those who did complete web checkout showed 170% higher renewal retention — not because the product experience improved, but because users forgot they had subscribed. Cancellation friction became a retention lever by accident.
This creates uncomfortable trade-offs. App store subscription management is imperfect, but users know where to go. Once billing splits across Apple, Google, Stripe, Paddle, or RevenueCat Billing-backed web flows, some customers manage subscriptions easily. Others do not. The result: lower voluntary churn alongside higher support load, refund requests, chargebacks, and long-tail frustration.
For apps on Apple's Small Business Program paying 15%, the net improvement compresses quickly. Stripe processing costs roughly 2.9% + 30¢ per transaction, plus 0.7% for Stripe Billing, plus optional tax tooling. On a $5 weekly subscription, processor fees alone approach 9%. Add Managed Payments fees, engineering time, experimentation overhead, support cost, revenue leakage from lower conversion, and dispute loss — and the 12-point commission savings shrink to near zero.
- Large U.S. user base with high ARPU and strong paid acquisition economics
- Existing web business where cross-platform account behavior is already the norm
- Sophisticated experimentation infrastructure capable of tracking full-funnel conversion, not just completed purchase margin
- Support capacity to absorb permanent billing fragmentation across platforms
- Merchant-of-record tools like Stripe Managed Payments or Paddle that handle tax and compliance complexity
Freemium Adds Another Layer
Meanwhile, freemium models continue to reshape monetization strategy expectations. Hard paywalls convert at roughly 10.7% median versus 2.1% for freemium, and generate 8× more revenue per install after 14 days. Yet top freemium apps — Duolingo, Slack, Strava — achieve 42–58% retention after one year, far outperforming average outcomes.
Freemium rewards exceptional execution. The difference between mediocre and great is enormous, and getting tier design right determines whether an app lands in the top bracket or struggles with sub-2% conversion indefinitely.
One recent case: transitioning from hard paywall to a multi-step freemium model (free access + 7-day trial of premium features) increased LTV by 75%. The business moved from excluding users upfront to growing rapidly through organic acquisition, but required significantly more sophistication in onboarding, upgrade triggers, and lifecycle marketing.
The strategic framework involves six steps:
- Define the goal — word-of-mouth, competitive differentiation, habit formation, network effects, or impact
- Map user success — identify the actions that predict retention and work backwards
- Choose freemium architecture — taster model (usage limits), split model (different features for different segments), or hybrid
- Map free vs. paid features — with clear strategic rationale and a "Bill of Rights" protecting core free value
- Determine trial strategy — including whether reverse trials (premium-first, then downgrade to free) fit the product
- Design upgrade triggers — contextual messaging, paywall timing, and visibility throughout the journey
Dynamic Paywalls and Smarter Testing
Paywall optimization has moved beyond static pricing screens. Tools like RevenueCat's new Paywall Rules feature allow developers to show or hide components conditionally based on user behavior, package selection, or custom variables — without new app releases. For example, displaying trial timelines only when trials are available, or swapping package options based on segment.
Testing volume matters. One running app used AI tools to scale creative testing from tens of concepts per month to over 400. This wasn't just about lowering CAC by finding winning ads faster — the rapid learning cycle fed product roadmap decisions directly. When hundreds of ad variations reveal what resonates, those insights inform not just acquisition but retention and monetization strategy.
Common paywall tests include:
- Trial duration — 3-day trials capture curiosity; 7+ day trials capture routine and habit formation
- Pricing tiers — apps offering 3 products vs. 2 see +44% conversion lift, especially with decoy pricing
- Visual hierarchy — motion graphics, personalization, and visible savings consistently outperform static designs
- CTA copy and color — small variations in button text ("Start Free Trial" vs. "Try Now") and decline button language ("No, I'll stay limited" vs. "Maybe later") yield measurable differences
- Feature lists vs. benefit summaries — testing whether "Advanced analytics" converts better than "See exactly what's holding your progress back"
The Real Cost Is Organizational
Most app-to-web discussions frame the decision as product or finance. In reality, it's a company decision. You need product and growth to redesign the purchase journey. Engineering to build routing, redemption, instrumentation, and edge-case handling. Support to answer "Where do I cancel?" forever. Lifecycle marketing to own web renewal and churn. Finance and ops to handle taxes, disputes, reconciliations. Legal or policy confidence on where external payments are permitted.
That's a lot of machinery to spin up to maybe improve unit economics.
The relevant question isn't "Can I avoid app store fees?" It's: "Am I capable of operating a better billing business than Apple or Google for this segment of users?"
For most teams, the honest answer is no. The app stores are expensive and frustrating, but operationally simple in ways people stop appreciating once they leave.
What This Means for Practitioners
Apple's Cal AI enforcement and the app-to-web economic reality point to the same lesson: platform power remains intact, and the path around it is not as straightforward as commission math suggests.
If you're evaluating web checkout, start with a constrained experiment in an eligible segment. Measure full-funnel conversion, not just net revenue on completed purchases. Track refunds, disputes, support tickets, cancellation confusion, and cohort quality. Check back one month, three months, 12 months after the experiment ends. Assume support burden will be higher and fee savings will compress.
If you're optimizing freemium, invest in tier design before paywall tactics. Define your Bill of Rights for what stays free, map features to strategic rationale, and test upgrade triggers relentlessly. The difference between median and top-decile freemium performance is execution, not luck.
And if you're navigating platform compliance, remember that Apple is still watching — even for viral, high-ARR apps. The rules may have loosened in the U.S., but enforcement remains active. Build your monetization strategy with that reality in mind.