Yesterday, at Startup Grind, an event series that’s aimed at new founders and people contemplating becoming entrepreneurs, we sat down with Joe Kraus, a partner for the last eight years with Google’s early-stage investing arm, GV.

Kraus, who’d earlier co-founded two companies — Excite and Jotspot — shared a range of founder-friendly advice, including what GV and Kraus in particular look for in founding teams (“irrational persistence”), along with some of the missteps that Kraus sees founders make, including “wanting to do everything at once.”

In fact, Kraus said there are three very specific steps that founders should do and in sequential order if they expect to raise seed, then Series A and Series B funding, beginning with “finding a product that serves a need in a market that matters.”

By “matters,” Kraus really meant “big.”

It may sound like a no-brainer, but Kraus suggested the GV sees plenty of founders who think they can win by dominating smaller markets. The problem, in his view: a mistake in a smaller market often means certain death, “whereas with a big market, you can make a mistake and the market carries you along.”

Kraus also advised focusing a lot less on “top-line growth” and instead on positive unit economics. (Despite the “obsession” of many founders to sell more to more people, often by throwing more product features into the mix, he proposed that VCs right now are far more interested in startups that make money off the sale of their products.)

And Kraus stressed the importance of  squeezing returns out of customer acquisition spend, be it through search engine optimization or content marketing or some other channel. Luckier startups that don’t figure out a winning strategy can “get caught in the tailwind of a market that’s growing and they grow with the market.” But you can guess what happens to the rest, he suggested.

More contrarian, perhaps, was Kraus’s advice to those who buy into the adage that founders learn from their mistakes. While Silicon Valley is known for embracing failure as a means to an end, Kraus pretty much called bullshit on the idea. The “story we tell ourselves,” that “you’ve learned more from your failures than your successes” is “dumb,” he said, explaining that while failure might build character, founders can’t learn much from failure other than what specific path not to take again.

For people looking to start a company, Kraus had this advice instead: Get a job at a successful company. While it might sound boring and corporate and like the last thing an aspiring founder might want to do, Kraus told the audience that you can learn a lot by making 10 minor decisions a day over a series of years — and observing the decisions of co-workers — if “combinatorially” they are part of a path that “yields success.”

Logging time first at a company that knows what it’s doing “gives you a pattern that you can follow in the future,” he said. Staring at the sea of young faces, he added: “It’s a far better path to learning [how to be successful] than shooting off on your own, trying something out and it not working.”

Read more: