This article originally appeared on The Wall Street Journal's blog on Feb. 14, 2014.
By Elizabeth Gerber
During the recent Olympics opening ceremony in Sochi, Russia– an iconic moment watched by millions around the world– five large synthetic snowflakes were bathed in light. Four of them electronically transformed into Olympic rings. The fifth failed to light up.
With more testing of the system ahead of time, it’s highly likely that this very public embarrassment could have been averted. Had the technicians probed their system in advance, they might have traded many private failures out of the spotlight for the one enormous public failure they suffered when the world was watching.
Even though the snafu was efficiently edited out for Russian TV audiences, this moment provides a valuable lesson. It is important not only to the organizers of the 2018 Winter Olympics in South Korea and beyond, but also to the thousands of startups working to launch their new ventures today.
The lesson is that mistakes ahead of time can be fixed. Mistakes during the launch can be costly.
For 18 years, I have worked with startups across a variety of domains including medicine, technology and finance. The most valued piece of advice that is often the hardest for startups to follow but is most valued and critical to their success is the following: “Fail early. Fail often.”
Let’s break it down and start with part 1: “Fail early.”
Failing early increases the likelihood of success later when more people are watching and the cost of failure is much higher. It’s about testing your core assumptions about your value proposition, revenue streams and cost structures early on when the cost of each mistake is relatively low. This early learning can inform product and business models, which become harder to change overtime.
Consider Kevin Systrom, CEO and co-founder of Instagram, who was working in marketing at Nextstop when he had the initial idea for Instagram. His initial idea was to make an app that allowed his people to check in with friends, make plans, and post pictures. He tested the core idea of his product by giving his friends a simple app programmed in HTML 5.
Based on early testing, he learned that his core assumption that one product could do many things well was wrong. He decided to focus in on posting pictures – an area that his friends and the broader market seemed to really enjoy. Had Kevin not tested this core idea early on in the process when he had little to lose, he may not have learned that his product would be most successful by focusing on one feature.
Now on to part 2: “Fail often.”
Startups ability to innovate depends on rapid experimentation – developing and testing new products and services and improving existing ones. Over 20 years of research, Harvard Business School professor Stefan Thomke finds that early failure is desirable because it is generally cheaper and gets rid of unfavorable options before more resources are sunk into them.
For perspective, lead innovation consultancy IDEO, expects no more than a hit rate of 1 in 300. That is to say, they need to have 299 ideas that fail to get to one that works. The 299 failed ideas are critical to get to the one success. The 299 failed ideas are viewed as 299 data points to lead them to what works.
They focus on learning from the experiment, rather than fearing embarrassment.
Consider the case of first time entrepreneurs, Mert Iseri and Yuri Malina who were starting, Swipesense, a company to tackle hand hygiene compliance in hospitals. (Swipesense was part of Design for America and a finalist in the Wall Street Journal Startup of the Year.) They spent the first year developing a single concept – a personal gel dispenser that hung on the pant leg of scrubs – without doing rigorous testing of the concept with real users.
In the first five minutes of their first focus group with doctors and nurses, they learned that no one was willing to hang the dispenser on their scrubs: They thought it would fall off.
So after a year of working, Mert and Yuri, received feedback in five minutes that changed the direction of their product. Now one year later, through many more rounds of testing, they’ve developed a dispenser that nurses and doctors don’t just trust to stay on, but want to wear.
So while it’s easy to say fail early, fail often, let’s be honest. It’s not easy. Failing is not fun. IDEO partner Diego Rodriguez and d.school co-founder Bob Sutton like to say, “Failure sucks, but instructs.”
For startup teams, faster failing means faster learning, and a much better chance of growing into a successful company.
So as we continue to watch the Olympics until the final medal is awarded, let’s remember that we are seeing performances by athletes who have fallen many times during practices, learned from their mistakes, and gotten up to try again.
You don’t make it to the Olympics without failing early and failing often.
- Elizabeth Gerber is an assistant professor of mechincal engineering at Northwestern University.