Pixar president on why building a 'fail fast' culture is hard
The duality of failure, IBM's bet on cognitive computing, and the difference between a data officer and an analytics officer: The Data Mill reports.
Ed Catmull, president of Pixar Animation Studios and Walt Disney Animation Studios, has been quoted as saying, "Early on, all of our movies suck." And he assured the thousands of attendees at last week's Gartner Symposium/ITxpo in Orlando, Fla., that he wasn't being modest.
The criticism sounds harsh, but Catmull doesn't mean it to. Instead, he means that new ideas are rarely perfect, and he encourages his team to iterate -- and even fail -- in their pursuit of the next great story.
Failing is not new to IT. The Silicon Valley tech mantra is arguably "fail fast, fail often," a phrase that has practically become synonymous with innovation. But, as Catmull pointed out, building a "fail fast" culture in the workplace is tricky. "The problem is, there are two meanings of the word failure," he said.
The first meaning relies on a rational, almost scientific look at the concept -- the one companies refer to when they talk about failing fast. "We all know that when we've failed in the past, that we learn from those failures and that they have made us who we are," he said. It's the second meaning that complicates things. "It's the one we learned in school, which is if you fail a class or a test, it means you screwed up, you didn't work hard or you're stupid," he said. Governments fail, relationships fail, businesses fail -- and none of these failures is a cause for celebration.
The fail-fast snag lies in this duality, especially if CIOs don't recognize and address its complexity. "If we're not aware of it, [employees] will fall back to the negative meaning," Catmull said. What's the best thing CIOs can do? Follow Catmull's lead and create a safe space where employees are allowed to fail, which means protecting new, fragile ideas, as well as the employees who are working on them. "Protection means … allowing them to make mistakes and not judging them in the process, as they're trying to work things out," he said.
Take Pixar's Up, the well-loved movie about an aging widower, a young stowaway and their grand adventure. When Catmull's team first conceived the idea, the story was nothing like the one viewers have come to know. It was about a castle in the sky inhabited by people who were at war with people on the ground. "This version didn't work," Catmull said. Neither did the next version nor the version after that, until, well, the version that garnered five Academy Award nominations, including best picture.
"If you're able to at least realize academically that [failure is] two different things, then you realize that failure is a necessary consequence of doing something new," he said.
Is cognitive computing the next big thing?
Building a digital business, one that tightly integrates the physical world with the virtual, is often lauded as a competitive differentiator. But at some point, all businesses will be digital businesses, and CIOs will have to find a new edge. According to Virginia Rometty, CEO, chairman and president at IBM, they won't have to look too hard.
What's the next defining tech trend? Rometty is betting on cognitive computing. "These are systems that understand not just data, but unstructured data -- image, song, video, you name it," she said at Symposium. "But there are two more differentiators: They reason and they learn."
Rometty said systems like IBM Watson, which she described not as a super computer but as a set of cloud services composed of 28 engines and 50 technologies, are a step beyond analytics because they can learn and respond in real time. Cognitive computing is already influencing IBM customers and partners in industries like healthcare, insurance and manufacturing. "We [work with] a company building children's toys, a little dinosaur that takes on a different personality based on the child and how that child interacts [with it]," Rometty said.
Rather than supplant human workers, cognitive systems provide assistance. "Every industry has data doubling every year," Rometty said. "It is impossible for your professionals or you to keep up."
It's no secret that technology spend outside of IT is on the rise, but just how much spend is debatable, it seems. "IT now controls only 58% of enterprise technology spending," Peter Sondergaard, Gartner analyst, said in the opening keynote at Symposium. "And it will be half by 2017."
In a separate session, Chris Curran, chief technologist at PricewaterhouseCoopers, based in New York, suggested IT only controls about 32% of technology spending. He based his figure on the firm's recently published Digital IQ survey of 2,000 global business and IT leaders, who reported "68% of the technology budget is outside of IT," Curran said.
Despite the discrepancy, the bigger point remains the same. "We, as IT leaders, don't control the whole IT budget anymore," Curran said. "It's not about, as Peter said, an environment of control anymore." And that, according to Sondergaard, affects the CIO's relationship with just about every member of the C-suite.
CDO vs. CAO
Struggling to understand how a chief data officer (CDO) differs from a chief analytics officer (CAO)? At Symposium, Gartner analyst Debra Logan provided a clear rule of thumb: She described the CDO as an offensive role of exploiting and monetizing data, as well as a defensive role of data governance and data quality. The CAO, on the other hand, is an offensive role only, responsible for defining the analytic strategy of the business.
If that doesn't clear things up for you, don't worry. Logan predicts the two positions will eventually meld into one. "We believe these two are really going to come together, and the CDO will be the overall role," Logan said. "The business is a data office."
"For the CIOs, you've just been too passive for too long. You're not conceptualizing how important you are in your company. You have to be an active leader, demanding your seat at the table because your company needs you, and your CEO needs to be more paranoid about how fast technology changes." -- Jeffrey Immelt, chairman and CEO, General Electric
"CEOs expect to see digitally attributable revenue double over the next five years." -- Mark Raskino, analyst, Gartner
"Smart machines are going to fundamentally change the game, change the core DNA of every industry. So get with your executives and think, 'What does that mean to us?'" -- Jackie Fenn, analyst, Gartner
"We would suggest that in the case of something like smart machines … invest early," -- Tom Austin, analyst, Gartner
"If you give a good idea to a poor team, they will screw it up. And if you give a mediocre idea to a good team, they'll either fix it or they will throw it away and come up with something new." -- Ed Catmull, president, Pixar Animation Studios and Walt Disney Animation Studios
Welcome to The Data Mill, a weekly column devoted to all things data. Heard something newsy (or gossipy)? Email me or find me on Twitter @TT_Nicole.
A look at how one CIO is running his IT shop like a startup.