When Failure Leads to Innovation, and When It Doesn’t (Part Two: The Leader’s Job)

For a PDF of this article, click here. Henry Ford said that failure is the opportunity to begin again, more intelligently. But many of us begin again with no increase in intelligence. Or, we don’t get smarter because we won’t risk failure in the first place. We covered how organizations fail at failure in our last post. Here, we will look at how leaders create organizations that fail productively – a necessary step on the path to innovation. Innovator’s Sweet Spot Relative to “smart failure,” there are two dimensions for innovators to optimize: (1) rate of non-fatal failure (experimentation that does not kill the organization) and (2) amount learned from any given failure. 1) Increase the rate of non-fatal failure with small, fast steps. As you may know, Bloomberg L.P. is a vast business media empire, founded by Michael Bloomberg, now mayor of New York. In his pre-political days, he once was asked how his corporation managed to complete such large information technology projects. He replied that they were successful precisely because they did not undertake large projects; they undertook lots of little projects. At a time when huge IT projects were all the rage, this insight was piercing. The many-small-projects approach bestows powerful benefits, central to which are that failures are small and learning is frequent. Brilliant investor that he is, Bloomberg had found a way to minimize risk while maximizing outcomes. In our own work assisting clients with strategy planning and execution we’ve learned that grand and perfect strategic plans are a waste. To paraphrase a military adage, no strategic plan survives contact with reality. It’s better to get your grand plans 60% right and then start an execution cycle of rapid, small steps that makes you smarter, fast – adjusting and refining the plan as you get smarter. As long as you’re going to be wrong (most often you are), you might as well be wrong sooner, with lots of instructive failures in the short-term instead of a few terminal ones in the long-term. 2) Increase the amount learned from failure. When an employee takes an educated whack at a problem, and the problem remains unsolved, that employee and her boss are at a crossroads. One path is to deflect responsibility, such as claiming that she had no choice. Or that it didn’t happen. Or that it was someone else’s fault. The other, better path is to pick the bone clean, with the employee learning every possible lesson from the tuition paid. Better yet, the lesson gets spread and learning is celebrated so that everyone in the team, department, or organization goes to school on one person’s tuition. Innovation How-To for Leaders Leaders can improve their organizations’ performance on both dimensions – frequency of productive failure and amount learned per failure – with some reasonably simple straightforward techniques. Here are a few to consider.

• Nix any project that does not sharply define its intended outcome. Your team will never learn what works and what doesn’t unless they have spelled out in advance the result they’re aiming for. It may be impossible, in advance, to specify exactly how something will be done (especially if innovators are improvising!) but it’s generally quite possible to spell out the result you are after. Without a crystal clear target, too much after-the-fact rationalization creeps in and then everything is an alleged success and nobody learns anything.

So, for example, next time someone wants to re-organize a department, ask him exactly what outcomes he’d like to produce, what side-effects he’d like to avoid, and how he’ll know if he’s been successful. Press hard for precision (blog regulars will be familiar with the Bar Bet as a litmus test of clarity).

• Make learning – rather than performing – the first task. When entering new territory, assume that you and your people are smart enough to learn, but not smart enough already to know exactly what you’re learning (otherwise, it wouldn’t be “new territory”). Define the task as one of learning before you define it as performing.

Goal researchers have found that performance improves on difficult tasks if your initial aim is simply to learn how to perform the task. In other words, there’s a time when perfect performance isn’t the key; at first, the key is learning how to perform well, which is different from performing well.

Good management consultants always enter uncharted waters with a “discovery phase” of the project before the “performance phase.” In the beginning, only the discovery phase can be clearly spelled out. How could it be otherwise? You have to learn what you need to do before you try to do it.

• Watch your language. Call innovation projects “experiments,” or “learning pilots.” Make it clear from the onset that the point is to figure out what works; or at the very least, figure out what doesn’t work. You can’t just give people “permission to fail and learn.” That permission has to permeate your language.
• Limit Your Losses. Target initial efforts that won’t kill you if they fail; don’t bet the farm until you’ve bet a few acres. For example, if you have a theory that putting a design team and an engineering team under one boss will produce more marketable products, then try one project that does just that; don’t change the whole organization until you’ve lowered your risks by upping your knowledge. Likewise, break your big, slow, risky projects down into lots of fast, little ones. Take a clue from Bloomberg and decrease the impact of failures while increasing the speed of learning. In theory, this approach should take longer, but in reality it doesn’t; but how many mega-projects do you know that came in on time? On budget? How many weren’t disasters? Think: “fast, small, and low risk.”
• Banish happy talk. Demonstrate that you are looking for truth, not Prozac. And then don’t punish the truth-tellers. When Alan Mulally took over as Ford’s president and CEO in 2006 he apparently got fed up with the deflected lessons that dodged both learning and accountability. As Economist tells the story: “He asked managers to color-code their progress reports – ranging from green for good to red for troubled. At one early meeting he expressed astonishment at being confronted by a sea of green, even though the company had lost several billion dollars in the previous year. Ford’s recovery began only when he got his managers to admit that things weren’t entirely green.”

Incidentally, we have to wonder if part of the problem was that “green” had not been defined.

• Make “Aha!” and “Doh!” part of every progress brief. While you look to your subordinates for results, also look to them for learning. When people brief the boss (that’s you), they need to know that part of the way to get an “A” is to share discoveries. If all you get is happy talk, then prod them: “Surely not everything has gone well; what have you learned from the glitches?” Assume glitches and applaud learning.

Everyone wants to look good in front of the boss. Just change the rules a little so that looking good includes excavating negative experiences for lessons. Just like at school, make learning something to talk about and evaluate – in addition to results.

• Feed forward lessons learned. Don’t just capture lessons learned. Require that plans for new initiatives demonstrate how they are incorporating past learning. One of us (Wendi) did that with project managers whose “lessons learned” exercise had become a useless bureaucratic exercise. By requiring new projects to demonstrate use of lessons learned, the learned lessons became applied lessons, leading to consistently smarter, more innovative projects.
• Embrace DISproof before you embrace proof. The point of experimentation is to get smarter, not to be right. So rather than tasking your team to prove that an idea works, task them to disprove it instead. For example, if a vendor you love has a new “solution,” find where it fails instead of looking for evidence that it works. Scientific philosopher Karl Popper taught us that we get much smarter by trying to disconfirm our theories than by looking for cases where we are right. This is a big deal.

Rapid, ongoing innovation demands that leaders treat intellectual capital like any other capital: accumulate it, nurture it, and use it. Requisite to that game is the organizational capability for frequent, productive failure. And that kind of smart failure requires smart leadership.

NOTE: Okay, we’ve emphasized here one type of innovation, the type in which a specific problem needs to be solved. We obviously aren’t touching on serendipity-based innovation, in which prepared minds get lucky, as when Columbus found America, or Fleming found penicillin. There’s a lot to be said for noodling around with your eyes open. But that’s a different topic. Or perhaps it isn’t? What do you think?

Published: December 7, 2011