Silicon Valley’s “fail fast” mantra has become gospel in recent years. Its evangelists promise failure, defined as a series of product tests and tweaks, leads to success. They urge failing fast, failing forward and failing often. They pepper TED talks with failure confessionals, and let’s not forget FailCon, an entire conference dedicated to the concept. But when is failure a key to building organizational strength, and when is it just…failing?
Failing fast can obviously be helpful. When developing a product or service for market, it saves time, money and grief to know what won’t work. We want to know the wheels are coming off early and while still in the parking lot, rather than later on the open road. Failing fast can save resources, and lives.
As in middle school science, the method begins with a hypothesis, testing it, and noting what happens. The fast fail approach accepts small, repeated setbacks, rather than barreling ahead with an untested idea. Going for broke may leave you broke.
Design and consulting firm IDEO, for example, pioneered processes to systematize innovation. Steps include structured brainstorming and rapid prototyping, or whipping up a scale model of a product. Combined with field research, the process helps designers sprint through bad ideas to get to the best.
Common sense, right? But a lot of organizations never bother with testing. They come up with a “good idea” and go with it, ignoring potential obstacles. When the product fails, rather than suffer a manageable delay, the business goes belly up. In Fast Company, Astro Teller, the CEO of X (formerly Google X), emphasizes the difference between making something and making something that works. To get to the latter, he advocates for a culture that allows failure.
Controlling the Fall
When Teller’s group built a vehicle that had “basically mastered” highway driving, “We probably could have made a lot of money,” he told CNN. But, after determining that human drivers weren’t reliable, and that they needed a fully autonomous vehicle, engineers took what they learned back to the lab to redesign. Test, tweak, repeat.
In its advice to startups, incubator Y Combinator describes a type of testing as doing things that don’t scale. As an example, early on, Airbnb offered customers “professional’” photography of their listings to make them more appealing to renters. The listings and bookings improved, and founders learned more about customers. “The process was entirely unscalable yet proved essential in learning how to build a vibrant marketplace.”
Beyond design consulting firms, hardware companies are also helping businesses embrace the failing fast concept. Technology, including 3D printing and Computer Numerical Control (CNC) manufacturing, has improved rapid prototyping, speeding product development for health care, technology, and the automotive industry.
Taking the fear out of the idea requires a clear-eyed understanding of the risks needed to reach a solution. Among the ingredients of failing smart, Rita Gunther McGrath in Harvard Business Review advises entrepreneurs to 1) decide what success and failure looks like before launch, 2) contain the downside risk, 3) limit uncertainty, and 4) codify and share learning.
A method that generates data also helps good ideas break through opinion-based decisions. “The one thing that can trump a higher-up’s opinion is data, and repeated experimentation and failure lead to a lot of it,” says Stanford professor Baba Shiv.
It Still Gets Messy
Even with safeguards in place, failing fast poses dangers. Gunther McGrath writes, failures “…can waste money, destroy morale, infuriate customers, damage reputations, harm careers.” Failure also takes a psychic toll.
In Wired, Erika Hall writes that “many VCs don’t care whether any one particular investment enjoys long-term success. They only care that a percentage of companies in their portfolio nets them a high return.”
She warns that fail fast inherently yields mediocre results blunting “creative possibilities to only those articulated by the target users, leaving designers devising a faster horse (lame) rather than a flying car (rad).”
Hall recommends more rigorous ways to learn through applied research to uncover not just what users like, but how they use a solution. Innovation leaders would benefit from shifting from “fail fast” to “learn quickly and think well.”
