Treating a TestFlight like a production launch
When Joe Fabisevich left Twitter to go indie, he didn't start by building his link-saving app, Plinky, in secret. Instead, he treated his TestFlight beta like a full production launch.
For months, he ran a beta program that required real infrastructure. When he had to change Plinky's underlying data structure, he didn't just wipe the database—he built a complete export and import system so the 5% of users who were deeply attached to their links wouldn't lose them. "It was not worth it in the sense of the value of the time that I spent on it," Joe admits. But the effort paid dividends. When the app officially launched and users inevitably asked for an export feature, the system was already built and battle-tested. Treating the beta with the care of a live product meant he already had the infrastructure—and the user trust—to scale.
The $10,000 Daring Fireball email
While building Plinky, Joe took a three-month detour. Six months after ChatGPT launched, OpenAI released the GPT-3.5 API. Joe and a friend saw an arbitrage opportunity: there was no official ChatGPT iOS app yet. They built Short Circuit, a barebones wrapper, in just 19 days.
The app's success wasn't driven by complex marketing or paid ads. It came down to one email. Joe pitched the app to John Gruber, framing it not just as another AI tool, but as a specific solution for Apple users. Gruber posted it on Daring Fireball. "We started getting alerts from RevenueCat in our pocket," Joe remembers. "We made like $5,000 in the first two hours." The app pulled in $10,000 in a day and a half. The lesson was immediate and clear: "Definitely reach out to press, you idiot."
Why freemium is a trap for indie apps
When Plinky finally launched, Joe wanted to be generous. He offered a freemium tier that allowed users to save 50 links for free. His wife, a staff product marketer, warned him they would churn. She was right.
"I went back and forth on freemium versus trial, and I settled on freemium because I want to be generous. And now I regret it," Joe says. The problem wasn't just lost revenue; it was lost data. Freemium users who churned disappeared invisibly, giving Joe fewer "shots on goal" to understand exactly when and why someone decided the app was worth paying for. He eventually dropped the free limit to 10 links and is now shifting toward a standard free trial model, realizing that charging users sooner is the only way to accurately measure the app's real value.
The three-email formula for app sales
Plinky is now priced at $39.99 a year, and Joe has found that the single best way to drive revenue is through structured email marketing to his free users. He collects emails upfront via Sign in with Apple, giving him a direct line to his audience that doesn't rely on App Store release notes.
His wife gave him a strict rule for sales: nobody cares if the discount is less than 50%. So he runs 50% off sales, and he never relies on just one email. The playbook is always three touches: an announcement email, a reminder a few days later to non-openers, and a final 24-hour warning. "The last email usually drives the most people through the door," he explains. If the first email doesn't convert immediately, the strategy is working exactly as intended.
In the full episode, Joe also talks about how he organizes his launches using a "lookback calendar," why he open-sourced his database library Boutique, and how he watched an AI agent completely rewrite the Short Circuit codebase in just three and a half hours.