As the co-founder of Branch, I’ve raised a $2 million seed round, and grown the company from the offices of Dogpatch Labs, Obvious Corporation (during its phase of incubating products), and Betaworks — whose “studio” model is one of the more progressive in the technology industry. So while I am relatively young and inexperienced (and likely somewhat naive), I’ve been exposed to an array of opinions and approaches to product building during my short time in the startup world. Through conversations with entrepreneurs and investors, it has become clear to me that the popular model for funding early-stage companies is fundamentally flawed — and likely stifling innovation.

At the core of the problem is the way investors evaluate early-stage startups. Provided you have a talented team, your ability to raise a seed round (and the terms) largely depends on two things: your product and “vision.” Investors want to see what you’ve built (or will build), and hear how it capitalizes on or causes a “massive shift” in the way the world works. However, if you examine the births of the most successful consumer internet companies, you’ll quickly realize that initial product and “vision” are flawed criteria.

More often than not, founders, early employees, and investors of the largest tech companies will tell you that the product that took off was never part of (at least not fully) the original “plan.” Twitter spun out of Odeo, Instagram was a Burbn pivot, BuzzFeed evolved from one of many “for fun” experiments — the list goes on and on. Heck, even Facebook was originally Facemash.

By and large, innovative products aren’t strategically imagined ahead of time – they’re stumbled upon while experimenting on-the-go.

Furthermore, even when the game-changing idea is discovered, the founders and early employees tend to underestimate its magnitude and potential. They’ll tell you that they knew the product was special but did not realize how how special it was or what it would lead to next. The best example of this: Google was a graduate school project that Larry and Sergey tried to sell for $1 million. And there’s no way that Jack, Ev, and Biz could have been predicted from day one that Twitter, the SMS-based prototype, would eventually enable revolutions and reinvent news media. But despite initially linear growth, they knew they had something special on their hands because the product they had built delighted themselves and a core group of addicted users. In the early stages of a product’s life, all you can know for sure is whether or not you’ve built something that delights yourself and others. The rest — how many others it will delight, and in what ways — is typically felt out along the way.

In other words, “vision” is a popular industry narrative applied after-the-fact.

Despite these patterns, investors require conviction and grandiosity from budding entrepreneurs. As a result, most “dumb” and “small” ideas never make it to pitch meetings or even prototyping, and instead die in the heads of talented builders. Unless you’re someone like Kevin Rose, you can’t walk into an investor meeting and say, “We’re not really sure what we’re building,” or “We’re not really sure how this idea evolves and gets really big yet,” and expect to receive funding. Which is a shame, given that Facebook, Twitter, Airbnb, Dropbox and the like were all initially “dumb” and “small.”

If the startup ecosystem wants to increase the rate at which world-changing companies are created it needs an early-stage funding model that encourages ignorance and experimentation.

That being said, some conviction and purpose is necessary because unprincipled ideation, more often than not, leads to fatal fickleness. If you put talented product people in a room and say, “Go! Build anything you want,” chances are they’ll just spin their wheels.

Thus, early-stage investors should be looking for entrepeneurs who are thesis-driven and implementation-agnostic.

Founders should be married to a thesis about the world — not a specific product concept. For example, when Jonah Peretti founded BuzzFeed it was with the belief that social networks were fundamentally changing the way news and information spreads. The product itself came later, and after many, many experiments. Similarly, Evan Williams recently founded Medium because he believes that publishing tools have barely evolved since he founded Blogger over a decade ago, and it is still too hard for most people to get their ideas out into the world. Both Jonah and Evan started companies out of conviction in their theses, not the specific implementations that would capitalize on them. This approach provides the freedom to explore and experiment (and get things wrong), within the creative constraints of a guiding belief.

In summation, early-stage investors should stop looking for fleshed-out products and grand visions from budding entrepreneurs. Instead, they should start backing bold (even wacky) product people with novel theses about the world around them. Or as Ev says,

Starting a company is like landing on the shore of a deserted island… You don’t know how big the island is at first or what predators lie in wait.

So let’s start backing folks who have a knack for finding deserted islands — not the ones who have the island and its predators all mapped out.