weitere ...

think: act: “We cushion ourselves from failure

Instead of making big plans we should embrace the fact that we can’t avoid mistakes, says bestselling business author Tim Harford. He suggests that trial and error is the best way to avoid major failure.


  • November 2012

Think:Act: Which personality traits allow us to learn from failure and then come back?

Tim Harford: People, in general, are not good at this. We have certain psychological protective mechanisms that help us cushion ourselves from failure. But that is not bouncing back; that is saying, “It wasn’t really a failure; it wasn’t really my fault. Actually things are fine, they will improve by themselves, given a bit of time.” One typical characteristic is called hedonic editing. This is where you take something good that has happened and something bad that has happened, and you put the two of them together – a bit like a child who is eating sausages and sauerkraut. Even though they don’t like the sauerkraut, they like the sausages so maybe on average it is ok. But this way we don’t learn from mistakes.

What would be a better way?

TH: Pixar have this tradition called “plus-ing.” Every Pixar movie is one long series of feedback, iteration, and improvement. Rather than saying something is very good or something is very bad, they ask what could be changed in the name of improvement. They will say, “maybe the light in this scene could be more atmospheric; maybe the sun could be a bit lower and the shadows could be longer.” That is a route toward improvement. So, as individuals, we should be concentrating on problems and fixing them. Rather than asking if you are succeeding or failing, take a far more constructive approach and ask, “Who cares whether I am succeeding or failing – what could be better?”

Which types of individuals are good at learning from mistakes and making a comeback?

TH: Comeback types usually are very good at seeking out advice from more than one person. Just because you are getting a second opinion doesn’t mean that the second opinion is right. So, you want a third, a fourth, and a fifth opinion. If you have got people who are talking honestly to you, you are going to get good ideas about how to improve a situation.

To ask for open feedback sounds like a no-brainer.

TH: But for somebody at the top of a company, this is particularly difficult. Who wants to tell the CEO that he is an idiot? Who wants to tell him that he is wrong? When General David Petraeus was younger, one of his officers ordered him to give him a report card every month criticizing how he was doing. It had to be an order, otherwise Petraeus wouldn’t do it. That was a real lesson in this great general’s education. He realized that senior people need junior people to be able to confront them. It is not enough just to be a chief executive and say, “Hey, I am in listening mode; I am ready for your comments.” You need to really search hard to pull honest criticism out of the people who work for you.

So the challenge is not to avoid crises but to learn from them?

TH: The problems we face are so complex that mistakes are inevitable. But, for some reason, we like to create plans, structures, and strategies that are predicated on the idea that there won’t be any mistakes. Look at the eurozone. It was constructed with clauses that basically said there won’t be any bailouts and there won’t be a failure. We don’t have any contingency plans. So, when we actually get into trouble, suddenly the whole thing is far more difficult to solve than we thought. My general advice would be to recognize that you will be making mistakes and to have systems in place to detect the mistakes early and to correct them – whether that is a monitoring system or an open culture where people can offer feedback.

Consequently in your latest book you recommend trial and error as a strategy. This would turn failures and comebacks into everyday business.

TH: This is clearly the strategy of companies like Amazon and Google. Both are constantly running experiments. Amazon is experimenting with pricing and marketing deals. Google is constantly experimenting with new services like Google Wave, which was a disaster, and Gmail, which wasn’t. Marissa Mayer, Google’s former Vice President, said that she is happy if only 20 percent of the company’s products are a success.

But you can’t build, say, cars by churning out beta-versions of your product.

TH: No. But there are products where you can fix problems very, very quickly. A restaurant can try new dishes; if people don’t like them, the next day they can do different dishes. So, the restaurant can learn very quickly. It depends on what the cost of failure is. In many cases the cost of failure is quite small and the potential reward is huge. If you were an entrepreneur with a bit of money, you could open five or six prototypes for a chain of restaurants until you find the perfect one. Then you open a thousand restaurants with that prototype. That is one potential shape of returns where the downsides are small and the upsides are big. There, of course, you want lots and lots of experiments.

When are experiments not such a good idea?

TH: When the upsides are quite small and the downsides are potentially big. The classic example is nuclear power. You have limited upsides from nuclear power. It is obviously nice to have, but there are other ways to make power. There are tremendous potential downsides if things go wrong. Therefore, the safety standards have to be very rigorous and the willingness to experiment very limited. A more recent example is banking, which has been very experimental at developing new financial products. These products have a very small upside and make a little bit of extra money for the bank, and with leverage you can turn that into quite a lot of money. But, when things go wrong, the downside that they generate to the economy is huge.

So, in this case, an experiment went wrong and turned into a major crisis?

TH: Yes, and it was because the three rules of adapting were ignored. Number one: you want lots and lots of different experiments. Number two: the experiments need to be small, because you want to lim- it this downside risk. You want to wait until you have a successful result before you push out and expand. Number three: you want very good metrics that tell you whether the experiment has worked or not.

We would argue that there has been too much experimentation in the financial crisis, not too little.

TH: In finance we don’t have lots and lots of experiments. We actually had a few simple ideas to repackage derivatives of different kinds of risks. Also this was not a small experiment but a very large one, because everyone was basically doing the same thing. And we didn’t have good metrics for spotting failures early. A lot of these bets were 30-year bets. Actually, we are still waiting to find out whether the bets were good bets. So, these were absolutely the opposite of the kind of experiments we would want.

Were CEOs and politicians too ignorant to deal with the financial crisis properly?

TH: I think a culture of top-down decisions and so-called visionary leaders is part of what led to the crisis and made it more difficult to bounce back. The problems we evolved to deal with, the problems our ancestors would have faced on the Savannah, were relatively simple. The idea that you have a leader who is going to take control of the elephant hunt, and tell people who is going to jump out in front of the elephant, totally makes sense here. That leader shouldn’t be taking too many risks, and shouldn’t be doing lots of experiments, because the experiments have severe downsides. It is not a good idea to say, “the tiger may be friendly, let’s experiment and see.” The sorts of risks we face today are much less dangerous but far more complicated. So, that traditional model of leadership is not going to get you as far. Today you need a leader who says, “I don’t know; let’s experiment, let’s explore.”

So the visionary CEO is overrated. What about experts and consultants – can they help prevent crises and enable us to bounce back?

TH: The psychologist Phillip Tetlock asked experts to make forecasts about the future. Forecasts about election results, economic questions. He asked historians, political scientists, economists, psychologists, but also journalists, diplomats, and civil servants. After 18 years, Tetlock published his results and concluded that basically experts cannot forecast at all. They are really only a little bit better than chimpanzees. The world is just incredibly complicated. It doesn’t matter who you are; you won’t be able to forecast. The best you can do is to be humble, and expose yourself to lots of different ideas.