My app failed, then I changed one thing, and made $80K

Key Concepts

• Bootstrapping: Building a business with minimal external funding, relying on personal savings or early revenue.
• ICP (Ideal Customer Profile): The specific type of customer a product is designed for.
• LTD (Lifetime Deal): A one-time payment for perpetual access to a product or service.
• BYOK (Bring Your Own Keys): A model where users provide their own API keys for external services (e.g., AI models), incurring the usage costs themselves.
• Product-Market Fit: The degree to which a product satisfies a strong market demand.
• Canvas-first interaction: A non-linear user interface for AI, allowing branching conversations and spatial organization.
• Electron app: A framework for building cross-platform desktop applications using web technologies.
• Vercel AI SDK: A software development kit simplifying integration with various AI providers.

---

Prrenit's Journey: From Software Engineer to Successful App Builder

Prrenit, a builder from India, quit his $250,000/year software engineering job to pursue building his own apps. His initial attempt with an app called Rabbit Holes AI (later Canvas Mode) saw no success. However, a strategic pivot in its pricing model and architecture led to $80,000 in revenue within six months.

The Product: Canvas Mode (Formerly Rabbit Holes AI)

Problem Identification: Prrenit observed that linear AI chat applications like ChatGPT struggle with long conversations, often losing context. Users also found themselves juggling multiple AI tabs (e.g., Cloud for coding, GPT for architecture, Perplexity for real-time info), leading to a fragmented experience.

Solution: Canvas Mode is a desktop application designed for "Power AI users" that offers a "canvas-first way to interact with AI models." This allows users to branch off conversations, create new nodes, and organize their AI interactions spatially, overcoming the limitations of linear chat interfaces. The idea evolved from Prrenit's own frustration as a UX designer trying to consolidate his AI workflows.

The Initial Failure and Pricing Model Mistake

Prrenit initially launched Canvas Mode as a web app with a typical subscription/freemium model, offering a free trial with limited questions before requiring payment. This approach garnered little traction for several reasons:

1. Lack of Personal Brand: Prrenit did not have a large personal brand or a "build in public" strategy to generate initial interest and trust.
2. ICP Saturation: His Ideal Customer Profile (ICP) – Power AI users – were already paying for multiple AI model subscriptions, making them hesitant to add another recurring cost.

The Pivotal Change: One-Time Purchase and Desktop App

To address the lack of traction and align with the realities of bootstrapping, Prrenit made a significant pivot:

1. Shift to One-Time Purchase (Lifetime Deal - LTD): He moved away from subscriptions to a one-time purchase model.
2. Bring Your Own Keys (BYOK): To make the LTD sustainable and avoid incurring ongoing AI usage costs, the app was designed for users to "bring their own keys" for AI models. This offloaded the AI API costs directly to the users.
3. Desktop App Conversion: To eliminate server costs associated with a web app, Prrenit converted Canvas Mode into a desktop application (an Electron app). This allowed the app to run entirely on the user's computer, making it a "ship and forget" business model from a cost perspective.

This pivot led to immediate success, generating $80,000 in revenue in the last 6 months. Users who purchased the LTD were more serious, provided valuable feedback, and showed greater belief in the early-stage product.

When Lifetime Deals Make Sense

Prrenit strongly advocates for LTDs for almost every bootstrapped business. He argues that:

• Bootstrapped vs. VC-backed: Unlike VC-funded companies that can afford to delay product-market fit, bootstrapped businesses need to validate their ideas and generate revenue quickly.
• Cash Boost and Validation: LTDs provide a significant upfront cash injection, validate the product idea, and build a sizable, engaged user base willing to provide feedback.
• Morale Boost: For solo founders or small teams, the immediate revenue and user engagement from an LTD can be a huge morale booster, helping to avoid getting "in their own head" about what to build.
• Flexibility: Even if a business already has a subscription model, offering a limited LTD can be beneficial.

Tactical Pricing and Growth Strategies for LTDs

LTD Pricing Strategy:

• Tiered Pricing: Prrenit started with a low price point of $29 when he had under 100 users.
• Incremental Increases: Once the user base crossed 100, he increased the price to $49.
• Future Transition: With approximately 1,200 users currently, he plans to increase the LTD price further and eventually transition to a recurring model. This strategy allowed him to gather early feedback and validate demand at a lower entry point.

Growth Strategies:

1. Correct ICP Identification: Prrenit emphasized the importance of correctly identifying the ICP, which for Canvas Mode was "Power AI users."
2. Tailored Landing Page: The landing page copy was specifically tailored to the ICP, calling them out directly in the H1 and other sections.
3. Twitter Engagement: Growth involved actively participating in Twitter conversations related to AI interactions and the new age of AI. Instead of direct product promotion, Prrenit "hinted towards something related to my product," engaging organically.

Business Metrics and Technical Stack

Current Business Numbers:

• Gross Revenue: Approximately $80,000.
• Profit: Around $75,000 (after refunds).
• Paying Users: Approximately 200, with 60% being lifetime users and 40% on a one-time annual renewal for updates.
• Total Users: Around 1,200.
• Unique Website Visitors: 38,000 so far.

Technical Stack:

• AI as Co-founder: Prrenit leveraged AI tools extensively, including Cursor for coding.
• Landing Page: Built with Next.js and hosted on Vercel.
• Desktop App: An Electron app built with React and TypeScript.
• AI Integration: Utilizes the Vercel AI SDK, which provides a high level of abstraction, allowing Prrenit to focus on the frontend without worrying about integrating multiple AI providers.
• Hosting: Cloudflare for general hosting, Vercel for the Next.js landing page.

Costs and Margins:

• Due to the BYOK and desktop app model, operational costs are minimal.
• Monthly Costs: Approximately $50 for Vercel hosting (for the landing page) and $200 for a Cursor subscription (for ongoing feature development, not maintenance). The business essentially has almost zero server or AI API costs.

Key Takeaways and Advice

Prrenit's primary advice for founders, especially bootstrappers, is to "find product market fit as soon as possible." He reflects on his past mistake of "appealing to my peers" rather than focusing on solving a real market problem. His core philosophy is that an entrepreneur's purpose should be to build something that genuinely makes people's lives better, for which they will be willing to pay. This focus on delivering value and quickly validating it with paying customers is crucial for success.