Trending: Seattle becomes first U.S. city to permanently require sick leave for delivery and app-based workers
Venture capitalist Marc Andreessen speaks at TechCrunch Disrupt in San Francisco on Tuesday. (GeekWire photos)

SAN FRANCISCO — “Fail fast” is a well-known phrase in the startup world — or “fail forward” and “fail better” — used by many entrepreneurs and investors to describe the process of rapidly identifying bad ideas and iterations as a way to ultimately arrive at a successful solution.

Marc Andreessen isn’t a fan.

The venture capitalist spoke at TechCrunch Disrupt in San Francisco on Tuesday, touching on topics like A.I., his VC firm, and of course, Twitter.

Asked about how entrepreneurs should balance daily work with long-term goals, particularly when it comes to treating failure, Andreessen cited military strategies that revolve around separating goals and tactics from one another.

“It’s incredibly important to have a really vivid, clear idea of where to get in the long run that you stick to and are very sold on and that everyone agrees to — then, be very, very flexible on tactics,” he said. “The problem is, it’s a dichotomy, a contradiction. You have to think about the long term, but also day-to-day tactics.”


Andreessen said that “fail fast” makes sense for tactics — “if a tactic doesn’t work, find another tactic.” But not so much for longer-term goals.

“I think ‘fail fast’ is catastrophic if it is applied to strategy and goals,” he explained. “A lot of founders talk themselves out of what are going to be good ideas in the long run because they aren’t getting immediate traction.”

Andreessen said that the ability for today’s entrepreneurs to access wide-ranging and immediate metrics, while sometimes beneficial, can also be detrimental.

“It’s really easy to get distracted by and draw long-term conclusions from that short-term data,” he said. “You see companies that have gotten in real trouble over that. Conversely, you see companies that work for a very long time on something that people think is just completely nutty and they get heavily criticized along the way.”

This is something that Andreessen has touched on before.

With the ability to receive immediate feedback in today’s rapidly-improving technology world, should startups listen to initial reaction from clients and users? Or should they stick it out?

There are success stories for both strategies. But Andreessen said on Tuesday that one fact has held true throughout time.

“For all this whole thing about how everything is supposed to be speeding up, it still takes a decade or more to build something really significant in this world,” he said. “It still really does.”

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Job Listings on GeekWork

Software Engineer III – DevOpsNational Center for Atmospheric Research (NCAR)
Technology Services SpecialistGlobal Innovation Exchange, University of Washington
Find more jobs on GeekWork. Employers, post a job here.