How can learning from mistakes drive a team toward innovation and success?
by Carol Hildebrand, August 2013
The image is indelible: ORACLE TEAM USA’s US$10 million catamaran, known in America’s Cup parlance as an AC72, nose down in San Francisco Bay while sailors scrambled to save the boat. Just weeks after being launched, the enormous craft pitch-poled, digging the twin bows into the water and tipping the sailing platform perpendicular to the water’s surface, propped up by the rigid sail wing. Buffeted by 25-knot winds and the pressure of the ocean current, the sail wing snapped, turning the entire boat upside down. A strong ebb tide swept the boat out past the Golden Gate Bridge into heavy swells.
“There’s no question this is a setback. This will be a big test for our team,” Skipper Jimmy Spithill told media after the accident. “But I’ve seen these guys in a similar situation in the past campaign before we won the America’s Cup. A strong team will bounce back from it.”
And bounce back it did. Through extensive use of technology such as data and video analytics, the team was able to not only repair but also significantly improve the boat. A scant week after relaunching its vessel in early February, ORACLE TEAM USA was able to line up with fellow cup competitor Artemis Racing in a training session on the bay. “The capsize was a pretty shocking event when it happened, but it was also tremendously helpful in the learning it afforded us,” says Christoph Erbelding, senior FEA (finite element analyst), structural design, for ORACLE TEAM USA.
Every team will encounter failure at some point—whether in software development, product design, or, as in this case, elite athletes training for the highest prize in sailing. But according to Amy C. Edmondson, Novartis Professor of Leadership and Management at Harvard Business School, most teams fail to see failure as an opportunity.
“We have to stop the belief that the best performers never fail, because actually, the best performers are those that fail intelligently,” she says.
Instead of finger-pointing, resilient teams want to figure out what happened, and then act on their findings.
The question is, how do you build a strong team that not only bounces back, but also pulls innovation from failure?
In the case of ORACLE TEAM USA, several things worked in their favor. This is not the first time the team has faced a challenge. In November 2009, the carbon-fiber mast snapped off ORACLE TEAM USA’s trimaran during a test sail—a major setback that could have hobbled the team’s cup run. In response, the team got rid of the mast entirely by swapping out the boat’s traditional soft sails for a rigid wing sail—a radical design change that proved vital to ORACLE TEAM USA’s ultimate victory over Alinghi of Switzerland in the 33rd America’s Cup.
That kind of shared, in-the-trenches experience helps develop a sense of trust and confidence among team members, allowing them to function as a team rather than a set of individuals, says Scott Anthony, managing partner at Innosight, a global strategy and innovation consulting firm.
“There’s a reason that venture capitalists like to back entrepreneurial teams with startup experience,” he says. “To borrow from Mike Tyson, ‘everybody has a plan until they get punched in the face,’ and the reality is that you do get punched in the face. Venture capitalists like teams who have already demonstrated that they know a punch is coming, and that they know how to respond.”
The second critical success factor lies in the team’s intense desire to glean every piece of knowledge possible about the boat—be it analyzing sailing data or figuring out what caused the boat to flip. This sort of attitude tends to exist in teams that encourage people to look at failure as an opportunity—another hallmark of resilience, says Anthony. “Instead of finger-pointing, resilient teams want to figure out what happened, and then act on their findings.”
It also helps that ORACLE TEAM USA’s design and racing strategies are based on the team’s ability to analyze and draw insights from the enormous amounts of data it collects, giving the team plenty of raw analytical material to work with. “The absolute core of our technical performance analysis is all based around collecting information while sailing, and processing it to provide useful feedback to the sailors and designers to improve the performance and design of the boat,” says Ian “Fresh” Burns, ORACLE TEAM USA’s design coordinator.
To provide that feedback, more than 300 sensors embedded throughout the boat collect a huge amount of performance data and transmit it to a server in the hull—about 3,000 variables running about 10 times a second. The sensors measure anything from the strain on the mast to the load on a particular rope; angle sensors on the wing sail monitor the effectiveness of each adjustment. The team also runs several video feeds, and takes still images of the sail wing every second. All told, this amounts to a gigabyte (GB) of raw data and 150 to 200 GB of video per day.
ORACLE TEAM USA uses the data in several ways. For example, it configures a light feed of about 150 key parameters and transmits the data in real time to the Oracle database on the performance chase boat. It also syncs the boat’s server to the team’s on-shore Oracle Exadata via a 4G connection, allowing team members to use their 80 GB of historical data for comparative analytics.
After the capsize, ORACLE TEAM USA deftly honed in on the aspects of the spill that contain what Edmondson calls intelligent failure, which she says occurs on the frontier of knowledge by teams that are trying to innovate and discover something new, while learning what doesn’t work. In the case of ORACLE TEAM USA, the team chose to comb through the boat’s reams of sensor data for clues as to what went wrong, putting less focus on other, less instructive, root causes.
For example, the accident was partially caused by San Francisco Bay’s wind and tide conditions—a freshening breeze combined with one of the strongest ebb tides of the year. This is what Edmondson calls preventable failure, and is best dealt with by learning what you can and then quickly moving on. Moreover, the sailors were still learning how to sail a brand-new, high-performance racing machine, which caused an atmosphere of complex failure, characterized by external conditions that cannot be fully controlled, regardless of how carefully a team prepares. “While you can be hypervigilant, this is another type of failure to let go of,” she says.
Total amount of photo and video footage collected every day by ORACLE TEAM USA to improve boat performance
Clearly, the sweet spot for innovation lay in the wealth of sensor data that measured how the boat weathered the strain of the capsize. The data generated by the capsize confirmed some structural assumptions while refuting others. For example, the boat stayed intact for almost three hours, until it got out into the very big swell, which backed up the team’s assumptions about the strength of the hull.
But sensor data also highlighted the limits of the stability of the boat, and whether or not team members had correctly gauged the loads being exerted on various components on the boat. Loads are typically on ropes, but they are also measured on standing riggings, shrouds, and stays; the strain exerted by a load translates directly into how much weight must be spent to make the supporting components stable.
According to Erbelding, the structural design principles that govern ORACLE TEAM USA’s boat are simple—make components strong and stable enough to meet the team’s needs, without adding any unnecessary weight. Indeed, the race’s strict weight limitations are a constant consideration.
The data related to the capsize helped Erbelding increase stability by allocating weight more accurately. “We got a lot of information on whether our load assumptions were verified or not—sometimes they were too extreme, and sometimes not high enough,” says Erbelding. This, in turn, allowed the team to shift weight around as it rebuilt the carbon-fiber hull. Carbon fiber is built up in layers, so the team was able to use the data to pinpoint where it needed to build up more layers to strengthen parts of the boat, and where the boat could be thinned down.
The result was a boat that maintained stability more efficiently, according to an article from Sailing World that reported on the revamped AC72.
“On the first run it was apparent that Oracle had solved some of the key issues that plagued 17—as the team’s AC72 is known—back in October, before the capsize shelved the big-boat sailing program for four months,” noted Sailing World’s Stuart Streuil. “Torsional stability issues highlighted by the flexing of the windward hull—‘racking,’ in America’s Cup parlance—were long gone. And Skipper James Spithill’s ability to keep the boat comfortably foiling was markedly improved.”
Erbelding acknowledges that the road to recovery wasn’t smooth, but he says that the group’s team culture enabled them to succeed,
“There are a lot of motivated, sophisticated, and skilled people working together,” he says. “If you need to change something, go ahead and change it—it’s inspiring, and once you get into it, it’s hard to go back to a more controlled environment.”
Carol Hildebrand writes about technology, business, and sports from Wellesley, Massachusetts.
Do: Build a team culture characterized by mutual respect and trust. “People need to believe that they can speak up honestly and without fear about large and small problems that they see, mistakes they may have made, and questions they have,” says Amy C. Edmondson, Novartis Professor of Leadership and Management at Harvard Business School.
Do: Build an accountability structure that supports a desire for excellence, not finger-pointing. “You want them to be afraid of the competition, not each other,” says Edmondson.
Do: Use failure to learn. “The ability to extract as much learning as possible from failure is what separates those who succeed from those who don’t,” says Scott Anthony, managing partner at Innosight, a global strategy and innovation consulting firm.
Do: Lead by example. A CEO who talks about where he or she messed up or celebrates the right kind of failure models behavior that shows failure as an important step to success, says Anthony.
Don’t: Wait forever to make the next attempt. Yes, you need to learn from your mistakes, but “a bias toward action really serves you well,” says Anthony.
Don’t: Try to be perfect. Ironically, many teams are so busy minimizing little failures that they don’t take the right kind of risk, or don’t take action on things that could contribute to a project’s ultimate failure, says Edmondson.
According to Amy C. Edmondson, Novartis Professor of Leadership and Management at Harvard Business School, keeping a sense of humor is a key part of the culture surrounding resilient teams. For some members of ORACLE TEAM USA, a postcapsize diversion appeared when grinder Shannon Falcone discovered the possibilities of the 2012 Red Bull Flugtag, a carnival-like event where teams launch homemade, human-powered “flying machines” off a 30-foot ramp into San Francisco’s McCovey Cove.
Hoping to help the team move past the accident, he asked ORACLE TEAM USA honchos for permission to build a flying machine out of the sail wing flotsam, and designed a minicatamaran topped with a glider-like apparatus. The project turned into a 10-day building spree for the team, complete with consultation from the design team. The apparatus didn’t fly very far, but it certainly accomplished its main goal as humorous bounce-back device.