The brutal truth for every founder chasing the next big consumer app? Mark Pincus, the mind behind Zynga, put it plainly on Lenny's Podcast: “The average app installs per user per month is zero.” Let that sink in. Zero. Not small, not hard, but effectively non-existent for your average user.

Pincus isn't just dropping stats to be a downer. He's laying bare the current consumer product battleground, especially as AI looms large. Everyone's asking if AI will be a new platform. His answer is sharp: “AI is not yet a new platform. It is an important technology.” Big difference. It's a powerful tool, sure, but it won't magically solve your biggest problem: getting users. Forget hoping for viral spread or word-of-mouth. As Pincus puts it, “that's hope strategy, not a belief strategy.”

So, what’s a founder to do when the distribution well has run dry, and your shiny new tech isn't an instant on-ramp? Pincus offers a few paths worth digging into.

Key Takeaways

  • AI is a technology, not a distribution platform. Don't expect large language models or other AI breakthroughs to automatically get your product in front of users.
  • Average monthly app installs per user is zero. The traditional hope for organic, viral consumer app growth is largely dead in today's saturated market.
  • Distribution must be a core product strategy, not an afterthought. From day one, your plan to acquire users needs to be as developed as your product features.
  • Explore "prosumer" models or services based on "free tokens." Pincus sees opportunity in products that leverage rapidly decreasing AI compute costs to offer previously uneconomical services.
  • Think 'agentic travel agent': Build services that deliver immense value by performing complex tasks for users, like a fully autonomous, personalized travel planner, enabled by cheap AI.

The Method: Build Distribution Into Your Product

Pincus doesn't mince words: you cannot build a consumer product today and then think about distribution. It's too late. The market is too crowded, and user acquisition costs are too high on traditional mobile and web channels. He argues your product's design must intrinsically link to how it gets into users' hands.

First, acknowledge the bleak reality: "The average app installs per user per month is zero." This means the classic consumer app playbook — build something cool, then market it — is broken. Pincus insists, “Distribution has to be part of your product and part of baked into the strategy deeply and and proven from the beginning.” This isn't about marketing tactics. It's about designing a product that solves the distribution problem through its very nature.

Where Pincus sees real opportunity is in what he calls "free tokens" – essentially, leveraging the rapidly decreasing cost of AI compute. Imagine if performing complex, agentic tasks for users became incredibly cheap. Suddenly, services that were once uneconomical become viable. He suggests building services that "re-imagine consumer services based on free tokens," calling it “such an interesting innovation zone.” His specific example? An "agentic travel agent." Today, a human travel agent is too expensive for most people's casual needs. But an AI-powered one, that could flawlessly handle complex bookings, personalize experiences, and adapt on the fly, for a fraction of the cost? “I think there should be an agentic travel agent,” Pincus states, identifying it as a place with "latent demand" previously constrained by economics. This isn't just about AI improving existing services; it's about making entirely new, valuable services possible by changing the cost structure.

Where This Breaks Down

Pincus's insight into "free tokens" and "agentic" services is compelling, but it relies on a few core assumptions that might not hold for every founder. First, it requires a deep understanding of current AI capabilities and, crucially, future cost curves. Predicting which AI "tokens" will become truly "free" and when is a gamble. Second, while the idea of an "agentic travel agent" is clear, applying this to other sectors requires significant creativity to identify "latent demand" that was previously uneconomical. It's a vision of what to look for, but not a step-by-step guide on how to find that killer, distribution-integrated product. Founders still need to do the heavy lifting of market discovery and product-market fit validation for these new categories.

What to Do With This

Forget your current product idea for an hour. Instead, identify three complex, time-consuming tasks that a "prosumer" (someone who performs professional-level tasks for themselves or their small business) in your target market currently struggles with. Next, brainstorm how AI, specifically its rapidly dropping compute cost, could automate 80% of these tasks, making a previously uneconomical 'agentic service' viable. Finally, map out how this newly affordable service would inherently distribute itself – perhaps by integrating directly into existing workflows, offering a dramatically superior value proposition, or leveraging unique data assets. Don't just build the service; build the service with its distribution strategy baked in.