[MUSIC] And in this session we're going to talk about failure and success, learning from failures and successes. Before we plunge into the material, Dadi, my friend is quite modest, but I think I can mention this just to establish his credentials. He became Executive Vice President I think in 2008, is that right? In the five years that he was Executive Vice President of Intel, a large corporation, Intel sales went from 35 billion to $50 billion annually. Dadi himself managed 35,000 people worldwide, a third of Intel's work force. And he also acquired some startups worth some $2.5 billion. So he knows the startup scene from the other side, from the side of an acquirer, and not just the side of an innovator. So Dadi, welcome, welcome back. >> Thank you. >> Add let's begin with learning from success and failure. Those of us who teach management, we can learn a great deal from the military. Sometimes we think of the military as not very well run. But actually most militaries, in my experience, are actually very efficient and well run. And one of the things that the military does in general is the practice of debriefing. When they have an exercise, a training episode, an operation, even a battle, they always analyze it and draw important lessons for future improvement. Now, we know that organizations and startups, they sometimes debrief when there's a failure. Why did this go wrong? What did we mess up? But in conversations with you, you've made the point that you need to debrief also when you have a success. You need to identify and pinpoint the reasons for success. Why did we succeed? So Dadi, can you tell us a bit about your own experiences and your views on debriefing and learning from failures and successes? What's the best way to debrief? >> This is one of the most crucial things that applies to innovators, to entrepreneurs, but also to normal people, normal human beings in the course of their life. In many cases, people like to call the brief a postmortem. I hate the word postmortem because it means that something died. But the postmortem is very much an exercise of learning that the medical science applies to understand what happened, what went wrong, and how to apply learning for saving lives in the future. I like the debrief, this is the term that the military uses. And a good debrief applies basically two parts, one what I call the practical part. In order to do a good debrief, you have to establish a way to collect data as you go, during the project or the operation and also analyze and summarize things as you go. Because by the time you do the debrief you have to have a lot of information and not to reinvent information. This is very critical, very important. And in some militaries I know they put it into a science of collecting computer data, video data, whatever is needed to analyze and reconstruct the events in the most accurate manner. The more important part is what are called the emotional part. People tend to look at the debrief as a potential of finding the person to punish. But this is not a judicial system. This is a learning experience. And in that the rules of this one are very basic. First, no punishment going to apply. Second one, everybody is equal, the big bosses, the lieutenants, and their worker base. Everyone has a way to put his criticism and to criticize anyone and to be able to be open. The big role of the generals or the bosses is to make sure that they open themselves to criticism and not scare everybody else. You'll be amazed how much more data in a much better way coming out that way because people like to defend themselves. And if they do, they will obscure the data. They'll have obscured the events, and you learn nothing from the results. The reason I mentioned earlier, Shlomo, that success is as important, and sometimes more important than failure, people tend to analyze failures. But they do not necessarily understand successes as well. And you'll see how much in the world, how many people done it once versus how many people have been able to do it again and again and again. Those that knew how to do it again and again again are probably those who understood why they have succeeded as well as they understood why they failed. And they've been able to establish and to keep the things that work well. >> I think this is a crucial point, innovation is a continuing, sustainable, sustained process. And unless you have a clear understanding of what works and how to do it again and again, once is not enough. Rarely is once is enough. So, Dadi, you've experienced a lot of great successes in your career. I mentioned some of the numbers. But you've also known really major failures, painful failures. And one of your successes actually came directly out of a huge failure. Can you tell us the story of how the Timna project, a failure, became a successful Centrino project? >> Yes, I think this is a critical case where the project failed. It was cancelled. We spent tens and hundreds of millions of dollars of a product that never got into the market which is a huge failure. Nevertheless, we were able to pick also what have done well on the failing project, which usually people tend to throw away everything else. To make the long story short, Timna was a product that we are trying to very much reduce the costs of a microprocessor to a PC. That was, in my view, one of the first what is now known as a system on the chip. That time it was not something that used to be done in the microprocessor business. And we have decided to apply an economical, the rule of the diminishing return. And what we realized at Intel, by pushing performance more and more, as a matter of fact, increased the area of the die of the microprocessor in a significant manner. And we realized that if you kind of stop at the knee of the curve, you gain huge amount of cost reduction. And we applied this one into this product. >> Dadi, can I just stop you for one second and show this little graph to our viewers, if we could zero in on that? So diminishing return simply means that as you invest more and more effort in something, the return you get back from it, we call it the marginal product, declines as you move along the curve. And your point is okay, so what if you move backward on the curve? By definition you get increasing marginal product. And that's the essence of your story. >> That's the essence of the story because what we realized that we save a lot of real estate. And the semiconductor's real estate costs a lot of money. We failed because we picked the wrong memory technology that failed to go through the cost reduction curve as we anticipated. And as a result, we did not meet the overall system cost goals. Nevertheless, we could have thrown the whole knowledge of what worked. What we have realized as part of the debrief was first, the product failed because of the memory technology, not because of the execution of the basic system on a chip. The second thing that we have realized, that we not only saved effort, and not only we saved on die area, we saved about 50% of the power for a negligible reduction in overall processor performance. The outcome of this one say, aha, we could take this idea into creating a high performing mobile part, which would not be able to rely on this unique memory technology. And we learned a lot of the lessons of what not to do as part of the team relating to this one. Eventually this became the Centrino. And what we will talk about later in a future lesson is that what were obvious to us, that you could apply to mobile, was not so obvious to the rest of Intel that was still in the culture of pushing performance at any cost. >> You still had to sell the idea even though- >> Absolutely. >> It seemed imminently logical.
