Stan Marchand, CEO of Rocapine, a Paris-based app publisher, is exploring the difference between genuinely valuable products and mere novelties in the context of vibe coding. Vibe coding is a new approach to app development that allows developers to describe their ideas in natural language rather than writing traditional code. This method has dramatically cut the cost of building an app, but Marchand argues that the real key to making money lies not in the code itself, but in the 20% of the process that involves human creativity, judgment, and strategic thinking about how to reach users. Despite the rise of vibe coding, Marchand points out that most of these apps do not generate meaningful revenue. The reason, he says, is that they often fail to connect with users. For an app to be successful, it must be marketable, able to convert users into paying customers, and retain them over time. These factors are measured by key metrics such as cost per install, conversion to paid users, and early user retention. Together, these metrics determine the return on ad spend (ROAS), which is crucial for evaluating the financial viability of an app. When comparing vibe-coded apps with traditionally developed apps, Marchand explains that the focus of evaluation is on the human elements of the app—its design, user experience, and strategic direction—rather than the code itself. As more developers use AI tools for app creation, the line between vibe-coded and human-built apps is blurring. However, Marchand warns that legal and structural issues can still arise. He recommends understanding the technology stack and its licensing, ensuring data is clean from the start, and making sure that the metrics used to evaluate success are verifiable. He notes that partnerships can be affected not by the AI-generated code itself, but by the difficulty in verifying the data involved. For app creators considering different business models, Marchand suggests aligning their choices with their personal goals. If immediate cash is the priority, a full acquisition might be the best route. Alternatively, publishing or revenue-sharing deals can provide long-term benefits with the support of partners in monetization and marketing. He cites an example where a publishing model led to rapid revenue before the app was eventually acquired. For vibe coders aiming to build and sell their own apps, Marchand recommends focusing on a niche they understand, moving quickly to launch, and pricing the app from the start. He advises against using generic features and urges creators to avoid "AI slop"—low-quality, generic content that can erode user trust. When working with publishers, he suggests bringing cohort data—information about user groups—rather than just listing app features.