Mark Pincus, the founder who built Zynga, has a sharp lesson for every founder drowning in their own brilliant, untested ideas: your instincts are often right about a problem, but your solutions are usually wrong. Pincus puts a number on it: “your instincts are right 95% of the time. Your ideas are wrong 75% or at best right 25% of the time.” That's a brutal reality check, and it's why his 'Proven, Better, New' framework isn't just theory; it's a battle-tested method for de-risking your product development. It’s how you fail for the right reasons, not the wrong ones.
The core tension Pincus addresses is that founders are drawn to novelty, but novelty is risky. Instead of betting everything on a hunch, he says, you should become a master of what already works. “Get your PhD in proven first,” he argues. “We haven't earned the right to innovate on the camera until we are the world's leading PhD on the best mobile cameras that already exist.” This isn't about being derivative; it's about building an unshakeable foundation before you ever attempt a moonshot.
His definition of "Better" also flips conventional wisdom on its head. Forget what you think is an improvement. "Better is something that 10 out of 10 of your existing the existing users of that product would say [__] yeah. 10 out of 10 not you. What you think is better is called new, right?" This demands an almost clinical detachment from your own biases, forcing you to focus on universally accepted, statistically verifiable improvements that existing users will switch for.
Only after nailing those two steps do you introduce the "New" – your truly novel concept, the "back of the box idea that's going to get people to download and try it." Pincus cites Words With Friends, where the novel element was its social integration. This structured approach, by isolating the high-risk innovation, allows you to fail fast and iterate on new ideas without bringing down the whole ship, because your underlying "Proven" and "Better" foundation is solid.
The Proven, Better, New Framework for Product Ideas
Proven: Make a list of things that are proven to be working already in the market, things people love about this sort of space. Look for what's the best of breed proven, whether it's Apple or Snapchat or Instagram, and copy those legally and with hopefully some taste. Be a master of the proven first; don't innovate on these aspects until you are a world's leading expert on existing solutions.
Better: Make something that is not just better, but that 10 out of 10 people will say 'f* yeah, I would switch at this is better.' This usually involves very small increments and innovations, such as making a product free or improving polish, which intense users will notice most. It should be statistically showable and universally approved by existing users.
New: Add a novel, new idea – the 'back of the box' concept that will get people to download and try it. Accept that this new idea is probably going to fail. If the Proven and Better parts are done correctly, the product has much better odds of succeeding, even if the New idea doesn't immediately hit, allowing for rapid iteration of 'New' ideas.
When This Works (and When It Doesn't)
This framework is a powerful tool for founders in competitive markets, especially consumer-facing apps where user expectations are sky-high. It works by letting you iterate quickly on high-risk innovations (the "New") because the core product experience (the "Proven" and "Better") is already robust and de-risked. Pincus designed it to help increase your odds of success by making sure you don't fail for basic, avoidable reasons.
However, this framework struggles when applied to truly groundbreaking, category-creating innovations. If you're building something where no direct "Proven" benchmark exists – think the early days of the internet, or the first truly smart device – rigidly following this model could stifle radical invention in favor of incrementalism. It’s best suited for spaces with established user behaviors, not for forging entirely new ones.
What to Do With This
This week, pick one product idea you're kicking around. Let's say it's an AI-powered personal finance tracker. Stop focusing on the AI. First, list out the Proven elements: what do people love about Mint, Personal Capital, or even basic banking apps? Identify their best-in-class budgeting UI, transaction categorization, or security protocols. Now, think Better: what small, obvious improvements would make 10/10 existing users switch? Maybe it's instant transaction syncing across all banks, or a truly intuitive goal-setting interface that requires zero friction. Only then, layer in your New AI idea – perhaps an AI that proactively suggests ways to save based on predicted spending patterns. Build your minimum viable product with robust Proven and Better features, allowing you to rapidly test and pivot the New component as needed.