September 2026 Summaries

RevenueCat’s Kotlin Multiplatform SDK replaced its CocoaPods-based iOS integration in version 3.0.0 with custom Gradle logic that builds Swift Package Manager targets from source and packages their Objective-C-compatible interfaces and static archives into Kotlin/Native klibs. Because Kotlin/Native cinterop requires Objective-C headers, Clang modules, and linkable libraries rather than Swift source, the implementation uses SwiftPM’s JSON package description, invokes `swift build` per Apple target, emits Objective-C headers, creates static archives with `libtool`, and generates module maps and `.def` files for cinterop. A source hash embedded in each `.def` file forces cinterop to rerun when Swift code changes, preventing stale Swift archives or bindings from being retained in cached klibs. Nearly empty facade modules such as `kn-core` then publish the generated bindings and compiled Swift code so client applications need only a Gradle dependency rather than CocoaPods or Xcode configuration. The system also handles release-build behavior, fixed toolchain versions, Swift resources through Compose Resources, dependencies and duplicated bindings across separate Swift targets, and Swift shims for APIs that cinterop cannot represent directly, illustrating both the flexibility and maintenance costs of integrating Swift packages into Kotlin Multiplatform before official SwiftPM import support matures.

WeWard, a French walking-rewards app launched in 2019, grew for four years through organic reach and PR rather than paid user acquisition, with a major television appearance generating roughly 300,000 downloads in minutes despite briefly crashing the app. Founder Yves Benchimol said this PR-led strategy worked in several European markets but failed initially in the United States, where the company lacked local recognition and a compelling media narrative. Its U.S. profile improved after tennis star Venus Williams became both an investor and partner, contributing to hiring credibility, lower acquisition costs, organic installs, media interest, and branded content. WeWard earns revenue through advertising, in-app purchases, a growing premium subscription, and affiliate commissions from brands including Nike, Adidas, Amazon, Expedia, and Booking.com, while declining partnerships it considers inconsistent with its wellness mission. Rather than optimizing for time spent in the app, the company measures success through users’ walking activity and requires product experiments to increase step counts; a recent feature that blocks selected social apps until a walking goal is met reportedly increased steps among users by 10%. Benchimol also discussed abandoning early Ethereum-based rewards after users preferred conventional loyalty points, the difficulty of ending unsuccessful initiatives, and B2B operations as a major challenge for the 80-person company operating across 29 countries.

Money-back guarantees are presented as a potentially useful alternative to free trials for subscription apps, particularly when used in web-to-app funnels and paid acquisition campaigns. Unlike free trials, upfront payments with refund protection can give advertising platforms such as Meta a clearer purchase signal and attract users with stronger intent, although they may reduce initial conversion rates and are unlikely to replace trials in every context. Effective guarantees should align with an app’s time-to-value, use clear and simple messaging, have practical refund processes, and be assessed through A/B testing based on net profit after refunds rather than conversion alone. Web payments offer the most control because app stores limit developers’ ability to manage refunds, especially on iOS. Guarantees may be most suitable for products with proven retention, meaningful paid advertising spend, and value that users can recognize within the refund window, while free trials may remain preferable for low-priced apps, products requiring long-term use to demonstrate value, markets without external payment options, or businesses still establishing product-market fit.

An analysis of more than 17,000 subscription apps from August 2025 through July 2026 finds that while high-volume apps commonly use short trials, longer trials are often associated with stronger conversion and first-renewal outcomes, particularly for monthly and annual plans. Trials of 17 days or more had a median conversion rate of 42.5%, compared with 25.5% for trials lasting four days or less, although the findings are correlational and may partly reflect the higher intent of users who remain engaged longer. Weekly plans showed modest benefits from 5–9-day trials, while monthly plans generally achieved the best balance with 10–16 days and annual plans saw conversion and renewal rise across longer trial periods. For AI apps, where free usage creates variable costs, 10–16-day monthly trials produced better renewal than shorter trials without the further conversion gains seen from 17–32-day trials. No-trial subscriptions generally had lower first-renewal rates for weekly and monthly plans, though profitability may still depend on pricing, acquisition costs, and purchase rates. Results also varied by region and category, suggesting that optimal trial length depends on how quickly users perceive value, whether repeated use builds a habit, local audience behavior, and the cost of offering free access; the analysis recommends testing durations against each app’s own revenue and retention data.

RevenueCat’s living guide explains how to adapt native iOS paywalls for Apple’s purported iPhone Duo foldable, which shifts between a compact outer display and a larger iPad-like inner display across closed, open, rotated, and partially folded positions. It recommends building and testing with Xcode 27.1 and the iPhone Duo simulator, presenting `PaywallView` in a system sheet so iOS can center it or move it away from the hinge, and carefully checking close-button placement, particularly when custom buttons may conflict with the camera area or vertical system toolbar. Apple’s related interface guidance emphasizes preserving the same paywall functionality across device states, avoiding major layout changes, keeping interactive elements and pricing details away from the fold, and using balanced layouts for full-screen presentations. RevenueCat’s paywall editor includes a Duo preview for testing screen shapes, safe areas, and orientations, while a final validation process should confirm that calls to action remain visible, package selections persist through folding, layouts avoid stretching, and the experience still works on standard iPhones and iPads.

RevenueCat’s App Growth Annual will take place on October 21 in New York and online as part of New York App Week, which begins October 19 and includes community events and The Shippies. The free event is aimed at app professionals and features five keynote presentations, 36 workshops, and virtual-only programming focused on subscription growth, monetization, user acquisition, retention, AI, creative production, experimentation, localization, team building, and investment readiness. Keynote speakers from companies including Tinder, Babbel, Burner, Tolan, and RevenueCat will discuss activation flows, contrarian business decisions, profitable churn, AI-driven cost structures, and RevenueCat’s upcoming product plans. In-person sessions offer practical exercises on channels such as Apple Ads, Meta, Google Ads, affiliate marketing, and organic content, as well as pricing, onboarding, segmentation, lifecycle messaging, and product research, while six recorded virtual workshops cover AI advertising, competitor analysis, conversion copy, Meta workflows, and feature development. Attendees can also participate in networking, an Expo Hall, a gamified virtual challenge, a live app-growth game show, and an afterparty, while keynote replays will later be available on YouTube.

Ambre, a recipe app developed by Gothenburg-based César Pinto Castillo and Sofia Larsson, evolved from César’s abandoned CloudKit experiment into a featured App Store product and a partnership combining engineering and design. César entered iOS development after accepting an app job before owning either an iPhone or a suitable Mac, while Sofia, a textile engineering graduate with no formal programming background, began by learning Swift Playgrounds and rewriting Ambre’s interface in SwiftUI. Her acceptance into Apple Entrepreneur Camp provided intensive technical, design, and business guidance that strengthened her confidence and helped establish her as a developer and co-founder. Although Ambre’s marketing has been modest, Sofia built an audience by emphasizing UI and visual design in a developer community often focused on code, while inventive promotions such as an App Clip-linked Swedish fish recipe reinforced the app’s personality. The discussion concludes that as AI makes software features and code easier to produce, product differentiation may depend increasingly on thoughtful user experience, quality assurance, design choices, and solving meaningful customer problems.

“Not enough usage” is a leading cause of voluntary subscription-app cancellations, often rivaling or exceeding price, but stated cancellation reasons can obscure deeper issues such as weak value perception, poor product fit, or a broken habit. The text argues that teams should distinguish usage churn from genuine pricing problems by defining what sufficient engagement looks like for their specific product, examining pre-cancellation activity, segmenting churners by usage levels, comparing price tiers, and combining behavioral data with user interviews. Rather than relying on discounts, which may reduce lifetime value without restoring engagement, app teams should identify where the habit loop failed: users may lack a trigger to return, fail to make the core action part of a routine, lose interest when rewards become repetitive, or conclude that the app does not deliver meaningful outcomes or fit their lives. Different usage-drop patterns can help diagnose these failures, while conversations with churned users, support feedback, reviews, and surveys can explain their causes. The recommended approach is to improve activation, contextual reminders, routine-building, fresh rewards, visible progress, user investment, and audience qualification, recognizing that some churn reflects an unavoidable mismatch rather than a problem to solve.

RevenueCat’s Android Purchases SDK supports several subscription-purchase architectures that independently determine who renders the paywall and who completes the transaction, ranging from fully custom interfaces using PurchaseParams to RevenueCat-hosted paywalls with optional custom billing logic. Custom paywalls fetch offerings and packages, display localized product pricing, and rely on Google Play’s returned subscription options to determine trial and introductory-offer eligibility, while purchase callbacks and coroutine APIs distinguish user cancellations from actual errors and provide updated CustomerInfo for entitlement-based access control. Subscription plan changes require an old subscription ID and a deliberately selected StoreReplacementMode, as defaults such as WITHOUT_PRORATION can cause upgrades to take effect immediately without charging the user until the existing term ends; modes differ in timing, proration, billing-date treatment, store support, and suitability for upgrades or downgrades. RevenueCat’s Paywall UI can be customized through options, listeners, activity launchers, gating callbacks, custom variables, and result handling, while PaywallPurchaseLogic enables an app to use its own billing implementation but requires it to acknowledge purchases within Google Play’s deadline and handle completion logic that RevenueCat would otherwise manage. The guidance emphasizes selecting defaults intentionally, correctly handling deferred plan changes and entitlement status, and treating a cancelled checkout as a user decision rather than a purchase failure.

Shipaton participants have two weeks remaining to target seven prize categories that emphasize market validation, growth, distribution, and audience-specific product design. Opportunities include building a Stripe-based RevenueCat web-to-app funnel for the $15,000 Funnel Vision Award, creating and monetizing an iOS app with Replit for the $15,000 Idea to Income Award, optimizing an Android app for Samsung devices and the Galaxy Store, and conducting a measurable growth experiment using the Layers SDK for the $15,000 Growth Loop Award. Three influencer-focused awards seek personalized wellness tools for Simone Sharice’s audience, satisfying nutrition assistance without calorie tracking for Abbey’s Kitchen’s audience, and AI-supported conversation rehearsal for new managers in the Leadership Heather category, with the influencer awards offering $20,000 first prizes. Entries must meet each category’s specific technical and submission requirements while avoiding unauthorized use of influencers’ identities or branding, and all projects require store listings, short demo videos, visual assets, and access to paid features for judges. Shipaton closes September 30, 2026, at 11:45 p.m. PDT, with developers advised to submit a minimal viable version early because app-store review may take several days.

AI is reshaping subscription apps but has not eliminated them, as app launches have risen sharply and established apps still capture most subscription revenue. The contrast between Chegg’s decline and Duolingo’s growth illustrates that products centered on one-off answers are vulnerable to free LLM chats, while apps that create durable value through structured workflows, accumulated user data, habit-building, specialized accuracy, emotional or community connection, and integration with physical-world data can remain competitive. Although AI-powered apps monetize and convert trials better than other apps, they generally have weaker long-term retention and higher refunds, suggesting that AI alone does not sustain subscriptions. The article argues that founders should test whether a general chatbot can replicate their core experience, then invest in differentiators that users cannot easily reproduce in a chat interface, communicate outcomes rather than merely AI features, avoid competing principally on price, and use LLM ecosystems both as competitors and potential distribution channels.

BoldVoice found that aggressive paywall tactics increased trial starts but produced disproportionately high refunds, leading the company to evaluate experiments by mature net revenue per exposed user after refunds rather than early conversion metrics. The company emphasizes continuous user research by requiring every employee to conduct interviews, involving heavy users in feature development, and collecting cancellation feedback to offset the bias toward satisfied respondents. Since launching in 2021, BoldVoice has charged from the outset and maintained a paid, professional-focused model rather than adopting freemium strategies designed for broader consumer products, offering limited free onboarding content followed by a trial. Its free Accent Oracle speech tool generated substantial attention after an investor shared it on Threads, illustrating how existing viral momentum can be improved with product features while not necessarily yielding high-intent customers. With roughly $10 million in annual recurring revenue and a team of about 10, BoldVoice attributes its efficiency to autonomous hiring, limited meetings, annual-plan economics, broad use of AI tools, and a specialized pronunciation-feedback product that it believes general-purpose AI models do not yet replicate.

RevenueCat has expanded Rico from an advisory assistant into a tool that can, with user approval, directly manage app monetization configurations and store products, including subscriptions in App Store Connect and Google Play. Rico can create and edit products, Offerings, Packages, Entitlements, paywalls, and experiments, handling tasks such as building variants, testing prices, and starting, pausing, resuming, or stopping tests. All actions require approval and remain constrained by the user’s collaborator permissions, while project admins can set Rico to read-and-write with permission, read-only, or disabled. Rico is also available in RevenueCat Mobile 2.0 for iOS, where users can ask questions, share screenshots, approve actions, and generate paywalls, while Siri and App Shortcuts can retrieve key business metrics; an Android version is currently being tested.

Subscription apps often face their highest cancellation risk immediately after conversion, with many trial cancellations occurring within the first day and many subscribers failing to renew, largely because users have not yet experienced clear value from premium access. The post-subscription gap arises when apps stop communicating after purchase, leaving subscribers unaware of unlocked features or unsure how to use them, which can worsen buyer’s remorse and subscription fatigue. The recommended approach is to re-onboard subscribers according to their status and guide them quickly toward a concrete “premium win,” an action or outcome they could not achieve on the free tier. A focused first-week journey can welcome users, explain premium benefits, prompt a meaningful first action, celebrate progress, and personalize later feature recommendations. Longer-term retention messaging can reinforce value through relevant updates, recommendations, progress reports, unused-feature spotlights, and identity-based achievements, while avoiding excessive interruptions for already engaged users. Teams are encouraged to treat conversion as the start of a retention relationship and measure early cancellations and premium-feature adoption to improve the subscriber experience.

RevenueCat Funnels enables app developers to create and host web-to-app purchase journeys for traffic from paid ads and creator partnerships, using existing RevenueCat products and entitlements to measure conversion from click to purchase. Recent updates include an AI Editor that generates and revises funnel designs from natural-language prompts, screenshots, or specifications; configurable bottom sheets and loader components for plan comparisons, recommendations, and timed transitions; and Firebase Authentication integration that lets visitors sign in or create accounts within the funnel. Developers can also add custom HTML, CSS, and JavaScript components for specialized interactions such as animations, calculators, charts, or WebGL experiences, while RevenueCat continues to manage purchase-critical elements. Together, these features aim to reduce the work required to test personalized web conversion flows and help campaign visitors move more directly from initial interest to purchase.

Subscription apps increasingly use hard paywalls that ask users to purchase or begin trials before experiencing meaningful value, a shift supported by stronger conversion rates but one that raises the stakes for onboarding, activation, and long-term retention. The passage argues that tactics such as winback offers, notifications, and pricing changes may address visible churn without resolving underlying gaps between marketing promises and users’ actual experiences. It recommends working backward from “must-have” users—those who would be very disappointed to lose the product—to identify the sequence of discoveries, emotions, and behaviors that turned them into loyal advocates. The Perfect Customer Loop exercise is presented as a way for teams to map this journey from advocacy back to first exposure, test their assumptions through research and retention data, and uncover the real value users receive, which may differ from the company’s intended value proposition. Activation is framed not as a single event or engagement metric but as an end-to-end story shaped by acquisition, expectations, pricing, onboarding, and early value delivery. By comparing the current user experience with the path followed by their strongest customers, teams can prioritize improvements to onboarding, activation, churn prevention, and product development that support sustainable growth rather than short-term conversion alone.

Paywalls V2 custom components let developers add self-contained HTML-based interactive content, such as Rive state-machine animations and Lottie vector animations, to server-driven RevenueCat paywalls when standard stacks, text, images, video, and buttons are insufficient. Bundles are uploaded as ZIP files containing an index.html file at the archive root along with all required scripts, styles, runtimes, fonts, and animation data, and they operate under strict security rules that prohibit inline scripts, runtime fetching, eval-like functions, and non-relative resources while keeping purchases and package selection in native paywall components. Rive implementations require a single-file WebAssembly runtime and embedded base64 animation bytes to avoid network requests, while Lottie uses a CSP-compatible lightweight renderer and embedded JSON animation data; both need responsive sizing, careful aspect-ratio framing, and pausing when the paywall is hidden. On Android, custom components require purchases-android 10.16.0 or later and Paywalls V2 but otherwise render through existing paywall presentation APIs, enabling dashboard-controlled animation updates without app releases. Developers must account for upload limits, archive validation, older-SDK fallbacks, and the larger footprint of Rive’s runtime, while using videos or native components where animation interactivity or custom rendering is unnecessary.

Developers aiming to meet the Shipaton deadline should prioritize getting a minimal, stable app publicly available in the US App Store or Google Play by September 30, allowing extra time for store availability and review delays. The guidance explains developer account requirements, release tracks, and Google Play’s 14-day closed-testing rule for newer personal accounts, which requires at least 12 opted-in testers. To reduce rejection risks, developers are advised to limit unfinished features, test thoroughly through TestFlight or Play testing tracks, prepare accurate store assets and privacy documentation, provide reviewer access credentials without two-factor authentication, and include clear instructions or demo materials for authentication, subscriptions, and purchases. It cautions against withdrawing or repeatedly updating submissions because changes can restart review timing, recommends checking submission details and policy links while waiting, and notes that expedited Apple reviews are intended for exceptional circumstances rather than competition deadlines. The suggested final-month strategy is to stop adding nonessential features, fix launch-blocking issues, submit an MVP as early as possible, and address monetization requirements carefully, while student entrants may instead qualify for the Next Gen category through a demo video and public open-source repository.

Colombian builder Peñalver developed Gurwi, an education app envisioned as a “Duolingo for everything,” after experiencing limited educational opportunities in La Guajira and teaching himself to code. He built an early Figma prototype that gained significant attention on TikTok, continued documenting his progress publicly despite limited institutional support, and received 10 million Colombian pesos from a Colombian tech entrepreneur who saw his content. Working with backend developer Martínez, he migrated Gurwi from FlutterFlow to Flutter and accelerated its planned launch to enter Shipaton, where they built and submitted the app shortly before the deadline. Peñalver promoted the project in Spanish and English through frequent build-in-public posts and creator collaborations that generated more than four million views, helping Gurwi win Shipaton 2025’s #BuildInPublic award and its $15,000 prize. Although visa and passport delays prevented Peñalver from attending the New York ceremony, Gurwi appeared in Times Square, and the resulting publicity contributed to more than $19,000 in user payments during the year, the formation of Gurwi LLC, and expanded monetization options. The team continues to add courses, explore business applications, and create personalized lessons, while Peñalver advises other participants to build products rooted in genuine motivations, tell compelling personal stories, make progress visible, and prioritize shipping useful features.

Built With Science grew Meta advertising to $100,000 in monthly profitable spend by directing about 90% of traffic to a personalized website quiz rather than the App Store, finding that the quiz improved conversion by educating prospects, addressing objections, increasing perceived value, and creating customized in-app training plans. Although co-founded with YouTuber Jeremy Ethier, the company delayed paid advertising for a year after its January 2025 launch to improve trial-to-paid conversion, retention, and churn, then built ads around testimonials, B-roll, and static creative instead of relying on his image. Its $189 annual plan gained adoption from roughly 60% to 75–80% after a pricing-page redesign displayed annual billing first and clarified the free-trial timeline, while adding extra explanatory content reduced conversions. A quiz-based “gym buddy” offer gives paired users a 15% discount, is selected by around 15% of trial starters, and appears to improve both order value and retention. The team treats its roughly 70% experiment failure rate as a source of learning, using funnel data, documented hypotheses, and iterative testing to improve results, while also developing features such as an AI chat coach and conducting in-app training studies that support product development and marketing.

RevenueCat has introduced custom components for Paywalls and web-to-app Funnels, enabling developers to embed self-contained HTML, CSS, and JavaScript for more tailored purchase experiences while keeping purchase-critical controls and checkout flows native. These components can support branded animations, animated backgrounds and icons, Lottie or Rive integrations, as well as accordions and expandable content for FAQs, feature explanations, and upgrade details. Developers can build components themselves or use coding agents with RevenueCat’s provided starter bundle, local preview tools, packaging utilities, and implementation guidance before uploading a ZIP file to the editor. The feature is designed for decorative and supplementary content, subject to security, file-size, structure, and supported SDK requirements, expanding customization without replacing RevenueCat’s native purchasing infrastructure.

Subscription apps experience their greatest churn at the first renewal, with roughly 42–65% of weekly, 39–58% of monthly, and 60–77% of annual subscribers failing to continue, while retention improves substantially after subsequent renewals. The central argument is that this outcome is usually determined before the billing date: users must reach meaningful value during an early activation window of about five days for weekly plans, two weeks for monthly plans, and 90 days for annual plans, rather than merely complete onboarding or open the app repeatedly. Weekly plans should drive immediate action and repeated core behavior, monthly plans should combine early value with ongoing habit reinforcement and transparent pre-renewal progress summaries, and annual plans require sustained engagement through the first three months and continued product relevance throughout the year. Companies can identify a First Renewal Predictor by comparing the early, high-quality actions of retained subscribers against those who churn, focusing on behaviors such as repeated core actions, meaningful feature use, or return patterns rather than raw session counts. The text concludes that poor first-renewal performance is primarily an activation and time-to-value problem, not necessarily a pricing, paywall, or win-back issue, and suggests measuring first-renewal rates by subscription duration through retention cohorts in RevenueCat.