Founder Playbook · Sub Club by RevenueCat
14 tactics from Charlie Cheever
Building Apps Faster: How AI and React Native are Changing the Game
Watch the full episode“In our report we showed that React Native apps actually monetize better than native apps... a lot of the VC funded apps are crossplatform from the beginning and kind of building a business not a hobby tend towards React Native.”
RevenueCat data: React Native apps monetize better — partly because serious builders pick it
RevenueCat's State of Subscription Apps report found React Native apps outperform native on monetization. The honest read is selection bias: indies stay iOS-Swift, while builders running React Native are usually shipping cross-platform from day one with a business mindset. Cross-platform from day one likely improves the revenue ceiling for new subscription apps.
“They had their sort of thing they'll say is like one man three platforms four months was how they built their first version because they weren't even planning to build an app... Paul on their team didn't even have React experience and he just learned on the fly.”
Bluesky shipped one engineer, three platforms, four months
Bluesky had no iOS or Android developers and a tight timeline when they realized they needed a flagship app for their federated protocol. One engineer, Paul, learned React on the fly and shipped across web, iOS, and Android in four months. They still only have about three people working on the app today — a benchmark for what a focused crew can ship.
“People review and they care about and they rate the outcome — is Craigslist helping me do what I set out to achieve? They're not rating every little interaction that you think needs to be perfect.”
Craigslist has 4.8 stars across 450K reviews — users rate outcomes, not polish
Craigslist's React Native app holds 4.8 stars across 450K reviews despite an obviously unpolished UI. Users rate whether the app helped them accomplish their goal, not micro-interactions. Many native apps with obsessive polish don't come close. Stop apologizing for shipping React Native — outcome beats fidelity.
“If you can't, you have to really know what you want when you're [building]. If it's to make it worth it, even if there is some relative difficulty of working with the tools, like it can easily be wiped out by a small increase in iteration speed.”
Iteration speed beats native polish — you can't validate what you can't ship
Charlie's designer-engineer friend abandoned a polished Swift app because slow iteration made it impossible to discover whether features actually worked. Unless you already know exactly what you're building, faster feedback loops dominate any quality advantage from native tooling. Discovery requires speed, not perfection.
“We just released a module not that long ago called like npm test flight where you can basically in an Expo project run npm test flight and it'll just walk you through all the steps and put it in TestFlight.”
TestFlight distribution — not coding — is now the bottleneck for vibe-coded apps
Vibe-coded apps run fine in a browser, but the real bottleneck is getting them onto the App Store. Provisioning profiles, distribution certificates, and Dun's numbers are knowledge most developers don't actually understand. Tooling that automates this last mile — like Expo's `npm test flight` — is what makes mobile distribution viable for non-experts.
“She was able to just fire up Bolt with Expo and like build an app just vibe coding that has their whole list of experts connected to that database and then has a whole sign up form... she was able to do that in a couple of hours.”
A non-dev vibe-coded a working app in hours that would have taken a week
The operator of a Seattle hybrid co-working/incubator space needed a custom check-in app that off-the-shelf tools didn't cover. Not a trained developer, she used Bolt with Expo to vibe-code a working app wired to a Supabase Postgres backend in a couple of hours. An expert dev would have taken a week. The build-vs-buy line for small operators just moved.
“Almost all new VC funded startups are using React Native versus building native, and huge apps — Spotify, Discord, NFL — you know tons of the biggest apps that people interact with on a daily basis are built in React Native.”
New VC-funded startups have flipped to React Native as the default
The framing that React Native is a compromise is outdated. The default for new venture-backed mobile apps has flipped, and the marquee logos (Spotify, Discord, NFL) prove it can carry products at scale. Native-first is now the unusual choice that needs justification.
“If the energy it requires to send somebody to RevenueCat or Expo or whatever is always less than the energy it would take to generate the tokens, like the model should just optimize towards that outcome.”
The energy-gradient: AI keeps recommending the same tools because it's cheaper than generating from scratch
Jacob's gradient-descent framing: models will route users to existing solutions whenever that's cheaper than generating equivalent code. Once you're the established 'relative minimum' for a problem (auth, payments, mobile delivery), the model keeps reinforcing the referral. New entrants only get a look where there's genuinely unexplored space.
“A big step forward for Bolt and Lovable was when they both took Supabase and made that sort of the way that we do a backend for you... AI generated code works better in general the more you kind of keep it on the rails.”
Bolt and Lovable hardwired Supabase because constrained AI generates better code
Vibe-coding platforms picked a single backend not just for setup convenience but because narrowing the AI's choices produces working software more reliably. The content angle for tool makers: write for the canonical stack the AI is being told to use, not the universe of alternatives. Generic 'works with everything' content gets nowhere with LLMs.
“He was like 'yeah pretty much everybody else works on the website.' I was like 'oh so the app must just be some sort of add-on here.' And he was like 'no no, 95% of our sales are through the app.'”
20-person team, 1.5 devs on mobile, 95% of revenue from mobile
Charlie recounts a 20-person dev team where one and a half engineers worked on the app while the rest built the website — yet 95% of sales came through the app. Teams habitually allocate dev resources by what's easy to hire (web), not where revenue actually happens (mobile). For subscription apps, hard-prioritizing mobile is usually the obvious unlock no one takes.
“A couple years ago engineers were sort of considered assets to tech companies and so people are trying to hire... if I can hire 100 people I'm worth $100 million... people are not thinking that way anymore.”
Engineers went from assets to liability — surface area is now a cost to minimize
Charlie describes a shift from headcount-as-value to headcount-as-risk. Companies with working native iOS and Android apps are migrating to React Native because maintaining two specialized teams is brittle, slow, and dangerous when a single specialist leaves. Engineering surface area is now a cost to minimize rather than a moat to grow.
“I think that we'll see a lot of winners win harder because people will just get started and not want to fight the AI that has just sort of been trained to like 'oh use this solution.'”
AI defaults mean winners win harder — get embedded as the default-recommended tool
LLMs default to whatever has the most training data, so Bolt and Lovable standardized on Supabase, and models reliably steer users to React, Expo, and a handful of incumbents. For distribution, getting embedded as the AI-recommended default compounds aggressively — fighting the model's defaults costs energy most users won't spend. New tools need to either crack the default or be unmissably superior.
“Here's your instructions — download the Expo Go app from the App Store, we're popping a QR code up on your screen, scan that QR code, bam, in less than 30, 40 seconds you now have your app on your phone.”
Vibe-coded apps reach a phone in 30 seconds via Expo Go QR scan
The onboarding flow for non-developers using Bolt or similar AI tools with Expo: scan a QR code and the app is on your phone in under a minute. This collapses the historical mobile distribution friction (Apple developer account, provisioning profiles, Xcode) into a single scan — opening app-building to non-technical founders the way Shopify opened e-commerce.
“If you just sort of say like go make me a website, like almost every model is going to default to doing React and sort of lean that direction and then probably tell you to go to Expo too.”
Software defaults to React because that's where the training data is
LLMs reinforce whatever has the most open-source examples. React wins websites by default; Expo wins mobile by default. For developer-tool SEO, this means tooling adjacent to the LLM-default stack gets surfaced for free in every 'how do I' answer the model generates — earning AI-default status is now a distribution channel.