The Starfish and the Spider | Rod Beckstrom | Talks at Google

Hi, everybody. Welcome to the Authors at Google series. My name is Joel Truer. I'm filling in today for Louis Monnier, who wanted to host today but is in France. Anyway, today I have the Honor of introducing Mr. Rod Beckstrom. Rod is a successful serial entrepreneur. He started his first company at the age of 24 when he was a student at Stanford. Sound familiar? He has also worked in philanthropy. He's an accomplished author and public speaker. He's worked in software, finance, micro finance, and the environmental movement. He is the chairman of the Global Peace Network, a trustee of environmental defense. He's currently the chief catalyst at a software company that he will not tell me anything about. He has helped create several organizations within Silicon Valley. Most notable to me, the Silicon Valley Social Venture, also the Environmental Markets Network. Today he's going to be talking about his new book, as you can see, called the Starfish and the Spider. Really interesting book. Draws its examples from Apache, the tribe of Native Americans, to Apache, the software package that we're all very familiar with. And the thing that I, as I was reading the book, found the most interesting was when you read the book, you will learn what Geronimo, the Apache, has in common with Mary Poppins. That was my favorite example from the book. Anyway, please welcome to the authors at Google Stage, Rod Beckstrom.

Thank you very much, Joel. Thank you so much. Well, it's a thrill and an honor to be here with all of you today. I love Google. It's an amazing company and certainly embodies a lot of the principles that we're going to talk about today. And I've got three simple goals that I have for our session today that I hope that you can leave with when we're done. And one is having a new model for looking at competitive landscapes and organizations and industries. And second is a model for looking at organizational structure and design, and thirdly, at looking at styles of management, leadership and collaboration. Okay, so those three models are what I'm hoping you're going to take away from the lessons that we're going to learn from the starfish and other creatures. I'm a co author of the Starfish and the Spider with Ori Brofman, and the book came out in October and we had the pleasure of recently launching the book in Italy, which was a lot of fun, and it's coming out in about 10 different languages and others underway, so it's a real thrill to be here. And in English, we have a saying, right? We talk about what is the mother of invention? What do we say? There's an expression. Something is a mother of invention. Necessity. We say necessity is the mother invention. But I actually believe that accidents are the mother of invention. Necessity can be helpful and necessity can drive things, but it's often accidents. And it's actually an accident that led to this new management theory being developed. And this is a new management theory that we're going to present today and a new universal model for looking at organizations. It's very, very simple. And by the end you'll remember the difference between starfish and spider organizations and be able to use that as part of your lexicon and your framework. And looking at. So what's the accident that led to this book? Well, it was a pretty big accident. I took this photo on 911 in New York City. I was there with my wife for a wedding that had occurred that weekend. And I was actually on the tarmac at LaGuardia when the first tower got hit on a flight that was headed to Dallas and then on to Mexico City for an environmental project I was working on actually for President Vicente Fox and the environmental minister Lichtinger in Mexico. And it was so shocking when the second tower got hit and I came out of the airplane and I was watching those towers burn, which was an incredibly emotional moment, I think for all of us, almost all of us. Was there anyone who wasn't, who was watching it as it unfolded on tv, as the events unfolded? Okay, almost everyone. So we know what an incredibly powerful moment that was. So for me this was a life changing moment. Completely changed my life. I'd done high tech stuff, I'd done environmental things, I'd done philanthropic things. But I'm sitting here watching these towers burn. And I had run Cat Software, which was a financial software company. We had 15 different customers in those towers. So we're a company with all these customers. I'd been there 200 or 300 times. So the feeling was just one of utter shock and awe. And I wanted to go help these people. I actually wanted to go get there and see if I could do something. And then I realized I can't do anything. I can't really go there and do anything. So I did the only thing we can do, which is pray. So I actually prayed for those people. And what I got in that experience was a sense that we have to do something in this world to improve our relations as human beings. In a sense, I was called to serve peace. And then I had a crazy idea. I said, I'm a business guy and I know I'm supposed to do work on Peace. So the first thing I did is I went to the Internet and I started searching all over, Googling terms, looking for peace organizations. And what do you think I found? Tens of thousands of peace organizations. Gave me an absolute headache. Didn't know what to do. So we had a crazy idea. So I teamed up with Ori and I had this crazy idea and I said, why don't we model a business network working for peace and model it on guess what network, Guess what network we chose to model our peace network on Al Qaeda. Al Qaeda, exactly. We made that strategic decision, which seemed pretty crazy at the time, but it ended up being just a really great strategy. And it was the experience of trying to figure out how Al Qaeda worked and building a network of CEOs that grew to over 4,000 CEOs taking part in explicit activities for peace across the Middle east, in India, Pakistan and the Mediterranean region as well. That led to this book and this work. So it was by accident that all this happened, and it was by trying to solve a problem, which is we wanted to study Al Qaeda, but there was no model. There were plenty of personal stories, histories, anecdotes. There was no model. So we had to develop one. That model is what you're going to have today for understanding human organizations and decentralized networks. So let's step forward and let's look at some interesting parallels or possible parallels. But what in the world, you know, does Wikipedia, and I assume virtually everyone here, since you're Googlers, you love technology. Does everyone here use Wikipedia? How many people use Wikipedia? How many people edit articles? Have edited articles? Well, see, that's higher. It's like 30% ratio here, and the average is like 5% out in the world. So Wikipedia, what does that have in common with Geronimo, with this guy, this gentleman in the center? This is Geronimo. What does he have to do with Wikipedia? Well, what does Wikipedia have to do with Craigslist? And I want to do a quick poll here. How many people have looked at Craigslist sometime in the last month? How many people have read classified advertisements in the newspaper in the last month? Oh, there's no hands up. Who wants to be in the newspaper business besides Rupert Murdoch, but actually he wants to be in the news business. So, I mean, Craigslist is a highly, highly disruptive company. So what does Craigslist have to do with Geronimo? What does that have to do with Skype? And what does Skype have to do with MySpace? What does MySpace have in common with Alcoholics Anonymous? The most successful addiction management program in the world that's helped millions of people. Does anyone here, by the way, know how many people are in aa? No problem. Nobody does. They don't know it's anonymous. They don't even keep membership lists. That's how they ensure perfect anonymity. You don't even have to register your chapter, just if you start one. But what is AA and that anonymity and that have to do with YouTube? And what do all those have to do with a guy in a cave, allegedly in a cave somewhere in Afghanistan or Pakistan? They all have one thing in common, one very simple and powerful thing in common. They're all decentralized networks, or leveraged decentralized networks. Every single one of those organizations, including Geronimo, who we'll come back to. But what is a decentralized network? Right? I mean, we have an intuitive sense. Yeah, decentralized network, that kind of makes sense. Right? And in the technology jargon, there's been studies of what decentralized networks are. But what does a decentralized network mean in this human social context? What is the architecture of a decentralized social network? How does it work? What holds it together? How does it start? How does it evolve? That's what we're going to look at today. But first, first let's go to biology. It was three years into our work. We were having dinner with Dr. Jane Lubchenco, a brilliant oceans ecologist. We're talking about what we're working on with the ideas. And the peace network was already building and unfolding and gaining steam by then. And by the way, the network had a real impact in India, Pakistan, where a group of only 11 people in that network staged a symbolic border crossing on September 13, 2003. The border had been shut, it was a war zone. They had a symbolic border crossing party. The border opened and it's been opened ever since. They also booked the first flight through a war zone in recorded history and got permission from both air forces to fly a jet through the shut border and had to develop a protocol for how you do that. How do you fly a private jet through a war zone safely with flight clearance from both air forces? That happened. They flew the jet through, it opened up the airways to all flights. So this network that we worked on, the peace network, actually did a lot of amazing things that we helped to catalyze what the CEOs brought to the table. And what they did was really remarkable. Anyway, back to biology. So we're having dinner with Jane Lubchenco and Jane says, you're talking about these networks, you're really talking about starfish. You should study the starfish. You've got to learn about the starfish. So we came up with the analogy of the starfish and the spider. It was actually for a presentation that I gave at Stanford Business School in May of 2004. So if we contrast these two, we're going to learn the new model, we're going to learn a new management model, a new organizational model. So let's start with the spider. Superficially, these two creatures look similar, right? They have a lot of arms or legs. They look very similar superficially. But internally their internal architecture is fundamentally dissimilar. And by studying that dissimilarity, we can learn a lot. So we'll start with the spider. Fuzzy little body, head on top, four legs on each side. If we take an adult spider and we cut off one leg, what do we get? Just get a crippled seven legged spider. You can Google it. Go to Google Images, do seven legged spider. You'll see a whole collection came up. You can even find six leggeds. I have not found a five legged spider living on Google Images yet, but it's a good research database, these topics. So a spider, if you cut off a leg, it just loses that leg if it's an adult. But what happens if you cut off the head of a spider? If you chop off the head of a spider, what's going to happen? Yeah, right. It's going to die. We know it's going to die. And the reason it's going to die is because it is a centralized organism and it cannot live without its central nervous system. And it has a hierarchy, has a very clear, has one stomach, has one digestive tract, it has one abdomen. So the spider is a centralized, hierarchical, clearly evolved organization. You cut off the head, it dies. That represents centralized organizations. We have studied centralized organizations in the west for 200 years. So when I was at the Beast School at Stanford and we studied organization theory, it was always with hierarchies, someone was in charge. There was a CEO or a chairman or managing director. Then there was VPs or EVPs and there was this really clear hierarchy, an organization. Whether it was a functional organization or geographic or matrix, whatever it was, there was a very clear organization. Those are centralized. But the world is being taken over by what new kinds of organization that are a lot more like the starfish. So this is a blue Linkia starfish here. It has five arms. Some starfish have up to 50 crown of thorns Starfish and others. And how many people here grew up within, say, 20 miles of the ocean or the beach? About half the class. So what happens if we cut off the arm of a starfish? Does anyone know? Yeah, it grows back. That's pretty amazing. Can grow back. The reason it can grow back is it's decentralized, okay. And it's able to regenerate. What we found out during this research process, in some species, like the blueling kea, if we cut off all five arms of the starfish, if you cut off all five arms, what do you get? Five new starfish. Each arm can regenerate an entire new creature. And the reason for that is that the starfish is so decentralized, does not have a single essential centralized organ. There's a nerve ring. There's a nerve ring that links the neural networks in each of the arms. But each arm has its own muscles, its own stomach. And that slit underneath has enough intelligence to right itself, to move around, to feed itself, and to regenerate an entire creature. This is a red Lynkia regenerating. And there's a whole collection I have online of photos that we found through Google image searches and others that shows how the remarkable regenerative capability of a starfish. And this has a lot to do with remarkable resilience and power of decentralized organization. And if you're going to go compete with them, they're working with very different rules, and we have to understand them to be successful in dealing with them. So let's go through history and take a look at what happens when starfish took on a spider. So let's jump back in time. So this is a story about a man in red tights, okay? Hernando Cortez. But. But don't think that this guy was a softy, okay? This guy was so tough. How many of you think you're really committed to what you do to your business or your causes? How many think you're really committed? Let's talk about this guy's commitment. This guy takes 500 men, sails all the way from Spain in these raunchy little wooden boats across a fundamentally uncharted ocean in search of gold and silver, okay? He arrives on shore. They disembark. Guess what he does next? He burns the boats. He burns the boats and he says, we're not going home without whatever silver and gold we can find. We're going to find it. We're going to get it. We're going to bring it home. Anybody have any questions? Okay, now there's no more exit door. So they march forward. And here's what they found. The Aztec empire. This is 500 men. And they find this empire with 15 million people. And they're going to take it on, right, and get all the gold and silver. Now, you look at this civilization. Does that look like a centralized civilization or decentralized? Yeah, it looks pretty centralized, doesn't it? It's well structured. Those are big pyramids. Someone's in charge here. This is, you know, this is ground zero of their civilization. So look, Cortes was from Spain. 1519, he arrived. Now, what was Spain like in 1519? Was it like a centralized society or decentralized? Sorry, it was a kingdom. Absolutely. It was a kingdom. It was a monarchy. And there's only one political system in power. Effectively, I'm trying to bring this down, so a little bit less feedback. So you had a kingdom and you had one church. You could have any religion you wanted to have, as long as it was Catholic. Okay? That was a society. So this guy understood centralized power. So how is he going to take down a centralized civilization? Easy. Find the leader, right? Negotiate with him. He got a little gold. Then he realized he had a lot more gold. He wanted more gold. Montezuma didn't want to give him more gold. So what did he do? Killed Montezuma. This should work here. Killed Montezuma, wiped out the royal family. Cut off the centralized water supply and roads coming into the city. Starved 240,000 people to death in three months. Got all the gold, all the silver, with 500 men, went back to Spain. Now, these are good venture capital returns, right? So, I mean, the Spanish. You're the Spanish king of the time or queen. What are you thinking? This is great. Let's do more. So the Spanish went down and took all of Latin America very quickly. They took out the Incas in exactly two years. Same model, two years. But they had a lot more gold and silver. Soon they swept all the way through the Americas, and they thought they were going to sweep the whole thing until they ran into this guy and his ilk, the Apache warriors. So this is Geronimo. Now, the Spaniards took on the Apaches in 1680. This is a photo of Geronimo in 1880. 200 years later, the Spanish failed to vanquish the Apaches. The Mexicans failed to vanquish the Apaches. And now the Americans are fighting the apaches. And in 1880, guess what percentage of the American army was focused on finding one man. 25%. 25% of the entire US army was chasing one man. Does this sound familiar, by the way? Bring it to present context. Okay, 25%. And what happened? What happened when they found him in 1886? They got him. They got him. It's great. We got Geronimo. Guess what happened? Nothing. They had a decentralized society. He wasn't the real leader. He was a charismatic leader. He was a guy with an attitude. He'd watch the Americans slaughter his wife and children. So he had a lot of energy in this battle and he was very creative and very smart. You could follow him if you wanted to. The Apaches had no formal hierarchy in their society. None. He was a non ton. So today we might follow Cory, tomorrow we might follow Sam, tomorrow we might follow Michael. Whoever you want to follow, you follow. Okay? That was their society. So the Americans went after him. They got him. No impact. In fact, the range of The Apaches in 1680, when the Spanish started was here. In 1880, it was here. They expanded their realm while fighting centralized powers. They did it. Now, how did America finally win? How did America finally get to a peace treaty in 1916, 30 years after catching this guy? 30 years. Here's how it happened. It was another accident. So the Americans set up the Bureau of Indian affairs, right, to correct the wrongs of the past. And someone at the BIA decided we have to give each American Indian a cow for every man, woman and child. And that'll give them some wealth and something to own and give them survival. Well, one guy was in charge of the Apaches, and he's like, excuse me, where are the cattle for the Apaches? And of course, the other people said, well, we're at war with the Apaches. They said, no, we're not at war with all the Apaches. We're just at war with most of the Apaches. What about the Apaches that aren't fighting us? Shouldn't we treat them well? And the BIA decided, yeah, we should treat them well. So imagine heading down towards the southwestern U.S. with 10,000 cattles to go meet a decentralized society. What's going to happen? Something has to happen, right? I mean, how do you give 10,000 cattle to a decentralized society? Well, someone had to decide. I'm the leader. I'll take those 10,000 cattle and I'll build the census system and we'll. We'll make sure we allocate this for you. Anyway, the gift of the cattle started to centralize their society, started to create a hierarchy to distribute those property rights and those assets and in the end, the people that also received those cattle started to settle down more. Right. Because they owned these cattle, they had to stay somewhere, graze them and move around. So what happened is it was the gift of the cattle that led the Apaches to peace. And it led us to one of the theories of how you can compete with decentralized intelligence entities, and that is to centralize them. If you have a decentralized competitor or opponent, it is often and usually in your interest to centralize them if you want to compete with them. So the cow approach is what we call centralizing the opponent. So that's the Apaches. But let's come back to the present. That's a long time ago. Let's come up and look at the music industry over the last seven, eight years. What changed in those intervening few hundred years? Well, one of the things that really changed was the Internet and ubiquitous communications itself. A completely decentralized network. Right. No one's in charge. It's committee process, human processes, totally decentralized network, making information ubiquitous and making communication virtually free and changing how humans can organize and collaborate. So the Internet came along and the world changed. Let's look at the music industry. So we have a new chart here, and this is a management chart that you can use at looking at any market space where we have centralized on the left and decentralized on the right. So imagine Attila the Hunt on the left, or top down, complete dictatorship on the left. And imagine complete randomness on the right extreme. And this is a continuum. We're going to look at the music industry over time, starting up at 1880, 1890, up to the present. So in 1890, the music industry was very decentralized. How many people in the room, by the way, play a musical instrument or have played a musical instrument? Okay, what instrument do you play? Clarinet, piano and guitar. You play the clarinet, piano and guitar. Okay, which is your favorite? Clarinet? Okay, clarinet. What's your name? Kurt. Kurt. So if it's 1880 and we want to hear Kurt play clarinet, what do we have to do? We have to go see. We have to see Kurt. Okay. There's no way he can record that music. Okay. There's no way he can record that clarinet. So the music industry was very decentralized. There was individual artists all over the world. You had to go hear them. They passed the music through sheet music so we could pass a song. But to hear that musician play, we had to go hear that musician play live. So then what happened? Well, then Edison and those guys invented the phonograph began to restructure the industry. So now we've got a record label that's going to do clarinet recordings. And Kurt's really good, so he gets the contract. Okay. But guess what? All those other clarinet players, that's 500 of them in New York State, they didn't get the contract. They're out of the game. So it's centralized. So the industry went from very decentralized up on the right. By 1945 in America, over 80% of music sales were done by 15 labels. By 2005, giant music companies owned 80% of global revenues. So total centralization of the industry, just a big move from decentralized down to that site. Then, right. Sean Fanning came along. Right. Sean was what, 18 when he started Napster? 18 or 19 as a freshman, wants to help his friends swap MP3 files. It's a real pain to do it. You gotta be a hacker to do it. Okay. You know, or coder. You gotta know how to use some things. And he's like, ah, let me make it really easy. So he made Napster so anybody could swap music for free. And it exploded, right? It exploded to over 40 million users. Now Napster's architecture was a little bit centralized though, right? Right. I mean, I think they had up to 10 million songs or more. And if we wanted to find, if I wanted to find Kurt's clarinet song, you know, Schupert classical recording of clarinet. I don't know if Schupert even did clarinet music, but I'm sure there's versions of that. If I wanted to get that, I had to go to Napster, go to the server, search on Schupert clarinet, find the song, and then it would link our PCs. Right. It had to go to the central servers and it would link. Right. So what did the music industry do? I mean, was this a problem for the music industry? Well, they hated these guys, okay? They hated Napster because their revenue started falling. Revenues fell 25% in the music industry because of this file swapping. And they took it all the way to the Supreme Court. They sued them, okay? There was a lawsuit, then it went to appellate court, Then it went to the Supreme Court. MGM vs Grokster. Finally. MGM won. They were so excited. Did it matter? Did it matter that MGO put a little peer to peer file swapping group out of business? Heck no. Didn't matter at all. It was a huge strategic mistake because Napster was trying to centralize and go legitimate at the time. And when Napster disappeared, what came up. What popped up next? Kazaa. Kazaa came out with a decentralized architecture. So that list of 10 million songs was not on the servers in San Francisco at Napster. It was where it was floating across all of our machines in the community. Right? They broke it into a little thing like encyclopedia with all those different volumes. Take 10 million songs and break it into a thousand little pieces. You got a million users online right now. Replicate each one of those volumes a thousand times, you ping each other. That's how the directory worked. Kazaa had no central servers, none except for ad serving, because they did serve up ads, even though it wasn't totally centralized, but it was managed centralized. So that's what Kazaa did. And then some people didn't like that. So they did Kazaa Lite, even more decentralized, took it out of their control. So when we attack a decentralized network, it tends to become more decentralized. That's the message. You have to be very careful in how you compete with these. All right, now let's look at your industry. We don't have a long time, but let's think about search for a second. You know, and how we used to search for information. You know, I used to go to the yellow pages 15 years ago, 10 years ago actually, you know, pretty pre 94 go to the yellow Pages now you go online. But if you look at you Google, you look at Yahoo, you look at different searches. Google strategy is a fascinating mixture because in many ways Yahoo seemed more decentralized than other approaches. You didn't have to go to all these different websites. They'd go and give you these listings and they had editors and they'd have their top listing links and you go to it, the link lists and you go search and find. But what did Google do? Google developed some algorithms, some great centralized search index algorithms that leverage the intelligence of the decentralized network that looked at who was pointing to who, that eventually looked at where traffic is. And all the magic soup that you put in the algorithms now. But think about it. It's where people click and what they look at that's driving the business, that's driving the search results. So it's decentralizing the process of ranking because the algorithms choose a rank, but it's all based on that external data that's out at the edge of the network that's leveraging tremendous power of decentralization. And it refers to the paradox of centralization, decentralization. Sometimes having a clean little algorithm or link in the middle can help a huge decentralized community come together. So you can use this tool on your own projects, whatever you're working on, in any area of the business or even outside the business. Now, all organizations aren't either a starfish or spider, right? So the truth is there's combo specials. I drive a blue Prius. I drove it here today from Palo Alto, you know. And is it an electric car? Well, sort of. Is it a gas car? Sort of. It's both. So not every organization is either a starfish or a spider. The starfish and the spider define the endpoints of the continuum. They define the end points. And then where you are, it depends on what your strategy is and how you look at your industry and how you look at your competitors. Let's give you a quick model, a quick five factor model for understanding the development of decentralized networks. There has to be an underlying energy to create a network that is then catalyzed by someone or some small group of people into creating a new type of network. They set the values and the ideology or, or bring values and ideology to the table. They then start circles that then can lead to network protocols unfolding. Let's step through each one of those quickly. The energy can be anything but primarily three categories. Passions. We have passions like Kurt loves music. Someone else wants to help save the world and the poor in Africa. Someone else wants to build great search algorithms. Someone else wants to build great image parsing routines or analyzing routines for search. We have different passions. We also have pain. You don't join AA for passion. You join AA because you've got a pain that you want to deal with. And then we also have physical needs. So there has to be a fundamental energy driver. If there's not an energy driver, a real network is not going to unfold. Then a catalyst has got to come along. Someone who's going to help organize this a network or give it some structure. Does anyone recognize this woman? It's a very important woman in America, in world history, and none of us know her. It's because she's a catalyst. It's Elizabeth Cady Stanton. She's the mother of the women's suffrage movement in America. She designed the network. She recruited Susan B. Anthony. We all know Susan B. Anthony because she was Miss External. She was the public speaker. She, she was the champion. Elizabeth Cady Stanton was the architect and was the catalyst who started the network. And she would have found Susan Binney Anthony or someone else to go do it. But Susan B. Anthony was an exceptional person. Who took the ball and ran with it. The lesson in that is catalysts can be very hard to see sometimes because they're not always boisterous or loud or exhibitionistic. Here's another catalyst, but this one you probably know, right? Who's this exactly? Craig Newmark. Very quiet guy, very brilliant guy. And he had values, which was he wanted to share what's going on in the community with other people. And then he kept moving that service to whatever people wanted, including exchanging products and services, meeting up, having events, whatever it is, whatever you want to buy, sell, trade, or connect to other people with. And he created this platform that the last time we heard was in 425 cities. That was a few months ago. So he's probably in 400, 475 by now. He has a set of values that are very different than Pierre Omidyar's values, who also had excellent values for creating this online marketplace. He valued anonymity. Pierre didn't. Pierre valued community and trust and building trust in the system. Different set of values, okay? But those values have driven that network, just like Pierre's have driven ebay's success. So you got a catalyst. What is a catalyst like? And are you a catalyst or are you more like a CEO in how you run your projects or how you run your group and how you work with others? The CEO, the traditional CEO, Right. We think about someone who's smart, decisive, driven. Think of someone like Steve Ballmer, okay, brilliant guy, but very driven, has the ideas of how he's going to run that organization. Decisive, rational. That's a traditional CEO. Powerful catalyst, tends to come across like a peer. Tends to work on trust, emotional intelligence. Okay, they can be brilliant too, but they have a different way of leading and guiding. And probably Larry and Sergey are great examples of catalysts and how they're leading this organization, but it's different. These are on the extremes as well. You got one on the centralized spider extreme, the CEO. And over on the decentralized extreme, you got the cat catalyst. And you can begin to think about every management decision that we make. In working with others, we're typically moving either towards centralization and taking power back to ourselves, or decentralization and collaborating with a group. Very different management styles. The Catalyst comes up with an ideology and a set of values, right? Like in this great organization, one of it seems to me one of the premier values is just fantastic search and incredibly quickly. It's. It's never fast enough. It's never good enough. You have to be the best on the planet. It's a set of values and many other values around that that led the organization to unfold, including not having too much structure and not being over managed. Shona Brown wrote a great book, Competing on the Edge, your SVP of business operations. And this book talks about how you need chaos in the organization and not to have too much structure. If your managers can handle everything they're doing, they'll have too much control. So you want to give them more to do than they can handle. So you keep it a little bit out of control. That's leveraging chaos theory or leveraging this notion of decentralization to further empower the organization broadly. It's a value that this organization has that drives innovation so well. The catalyst typically starts a circle, usually starts with a circle. An AA was the first chapter and the Woman Suffers Movement, it was a circle. Or in Wikipedia, it was the first set of articles in English and the first contributors. But you start a circle, you start a network, a group. Then if that's successful and you learn from it, you start to develop protocols for the network. Every network has a common set of protocols. Human network, there's rules on membership, right? MySpace, what's the membership rule? Anybody can do it. Anybody can do it. Okay, anybody can come in, you got to do legal activities, certain things you can't do. But anyone can come join and there's process of removal. If you abuse the system, you're thrown out. Every network has some kind of rules, whatever it may be. You have circles and peer groups you create in those. The network itself has rules. What's the connection between those pages or between those individuals, or between the working groups and circles? Who funds whom or is there any funding or what funds that network? And then what are the governance rules? And I'm looking more on the human and social side right now, but I. But the same applies in technology, right? You have a set of protocols around these, you have protocols in networks and how you design networks and how they work and how they unfold and the applications that you may build on top of networks. So that's a simple five factor model. Now how do you compete? There's certainly three major strategies for competing with decentralized networks. You can work at a level of ideology because the networks are driven by values and ideology and, and that's all about human psychology. So if you're fighting with Al Qaeda, are you going to win by arguing with them about their belief system? No, it's just going to lock them in. But you can flank them or you can Shift the social circumstances, focus on education, focus on microlending. Do the kind of good things that google.org is working on to change the social environment, to change the underlying problems. Okay, so you can work on ideology and the environment. Secondly, you can centralize them. We talked about the cow approach with the Apaches. Thirdly, you can decentralize yourself. You can decentralize yourself. So if you're the US military and you're fighting a decentralized terrorist opponent, you do more special operations in small groups of people in the field as opposed to large centralized congregations that feel safer but actually make bigger targets and then don't relate to society in the same way. And the generals leading this war realize that they're reinventing the entire US Military right now based on the third principle, decentralizing themselves with information, distribution of power in the field, et cetera. Or you look at your organization and you look at innovation, right? Looks like a lot of people here are probably in engineering. And you get a day, a week to work on what you want to do. That's decentralizing creativity in your organization because you get to do what you want to do that can contribute to the business. It's very empowering and it unleashes the power of all the individuals and engineers in the organization to contribute. So that's another example of decentralizing yourself in a fashion. And that's why I love the interview that was with Larry and Sergey and I can't remember if it was Fortune or Fortune. Someone asked them, what's your strategy for the next five years? And they said, we don't have one. We'll see what our smart people develop. And that's the truth. Right? You don't know exactly what's going to be the killer app or the killer hit product you come out with in three or five or six, seven years time. But it's going to come out of the smart people in this room and the rest of the organization or the others that join in the future. That's where it's going to come from. So those are the three key strategies. I was going to talk about projects in the slums. Microlending is a great tool. Microlending decentralizes banking for the poor. It decentralizes credit risk down to the borrowing circles of typically five people, usually five women, because you are co insuring each other. So microlending is basically making small business loans like $100 per person or 300 to the poor to run a business. But five of us would come together in a circle in microlending. And we have to create a group together and insure each other's loans. When we insure each other's loans, we're taking responsibility for the credit risk. We just decentralize it down to our circle away from a big fat bank that the poor might not pay to. So there's architecture that can be used here even to finance eloan and prosper. Prosper.com is an example of trying to decentralize the lending market in the U.S. other Stories this is a city in Kenya that's going to be built. I'm going to do a couple of slides on new work that I'm doing. I'm actually working on a next book that's going to be called Passion Networking. So what I do is really quick. So imagine this network architecture you've seen and we've talked about in decentralized networks. And now let's bring in the human component. So there's a research group across the hills here over in, I think Felton County, California, and they use these electronic equipment to like measure the human heart, right. Do you know that they can measure the human heart from 15ft away now they can pick up your EKGs. Now what does this have to do with networking? Well, hearts have a lot to do with networking, right? Because it's about our passions and what we choose to do and what groups we decide to join online or not join and not do. The basic notion of passion networking, and I'm going to skip this slide for sake of time, is how do we identify other people's passions within an organization or across a society and help bring together working groups around those passions to drive personal life change, social life change, or business change? I mean, even if you take the most esoteric interest in the room, like collecting 100-year-old butterflies from Africa, or take any esoteric interest someone might have, it's probably someone else in this organization of 10,000 people around the world that might be able to relate to that and feel passionate about it and want to share or exchange ideas where the Internet's going. Where I see the social networking domain going is in helping people to really link around their passions more deeply with whatever level of publicity or privacy that they want to have. Because privacy and intimacy becomes very important as well. And I think we've seen an explosion of the peer to peer transactions on the web, okay. And the one to many model like the Google search engine to, you know, a billion users. But what we're going to see next is circles coming together, social circles and the question is, how do we do that around people's passions, around their pain or their interests, and do that in a way that's meaningful to them and create these new spaces. We're starting to see a first new bunch of companies, you know, popping up, the Nings and the others trying to support this. And we're MySpace and Facebook or evolving. But I think there's going to be a real art and science here that's fascinating of how we literally network around our passions. So that's the overview. And if anyone needs to reach me, I'm rodxtrom.com, and I'd like to do now is answer any questions you've got on the framework, the starfish and spider framework again. And I should actually summarize. Let me summarize really quickly. Three things. We talked about a model for looking at competitive space. Like we looked at the music industry, but we could look at the search industry or the email industry or the golf industry or the rock core. We can use any industry and use that map centralized or decentralized. We talked about within the organization, how do we structure things on that same continuum. We can look at our own projects and activities and finally a management model of the catalyst versus the CEO and think about what are we and where do we want to be? Where do we want our projects to be? So those are the three tools that I hope you got out of the presentation. And I would like to take any questions.

Well, so my question is Mike or no Mic, because I didn't think of it until you used the example of the starfish that regenerates the whole rest of the starfish from one arm. So that made me think, what do the arms get out of each other? So the, you know, if the arm can live all by itself, why does it need the other arms? And I think the analogy is why in the decentralized organization, the cells need the other cells. Like, what does that really help?

Great, great question. And the first thing that I want to mention. I'm sorry, could you repeat the question? You want me to repeat the question? Yes. The question was if one arm of a starfish to that's going cut off, or all five that are cut off or 10 or 20 can regenerate a new starfish, why do they need each other? And why don't they just choose to break off? Guess what? In some species, they do break off voluntarily. Some starfish actually will reproduce by ripping themselves in half. Okay, they'll rip themselves in half and regenerate. It's one technique they have for reproduction. And so the truth is those arms don't need the other arms, except that it's. Guess what? It's the teamwork and collaboration. So, actually, let me tell you something interesting is how the starfish hunts for food. Okay? So you got five arms, right? And each one's got a sense of smell, and each one's got a sense of temperature and texture. They got all those soft, touchy things in the bottom, the starfish that not only stick to things, but they can feel things through those, right? So how do you choose? Let's say you're on the side of a pier, right? You're on this, like, wooden thing on a pier, and there's oysters above and below or whatever it is that's hanging on the side of that pier. How does a starfish choose which way to go? Maybe both arms are touching something. That's lunch. Okay? They communicate. They communicate in terms of the intensity of the signal of how good that lunch appears to be and how likely it is it can be won. And guess what? There's a tiebreaker. One arm is the tiebreaker. There's an alpha arm, okay? They don't know which it is, but there's an alpha arm that can be the tiebreaker. If the starfish is, like, locked between two arms, like, pulling in opposite directions, one can say, okay, we're going that way. And it's kind of like our teamwork, right? You know, trying to work with consensus works really well most of the time. But if we get completely deadlocked, we gotta get undeadlocked, right? And we gotta let someone be the tiebreaker, or someone's got to kind of surrender and go in a direction. But starfish are really fascinating. And of course, there's not that much known about them, I'm sure, compared to what we're going to learn over the coming years. But they're fascinating, fascinating creatures. Really interesting. Thank you.

Which organizations are well suited for centralized structures and which ones are well suited for decentralized.

Great. The question is, which organizations are well suited for centralized structures and. And which organizations are better suited for decentralized organizations? Let me give you an example. My daughter just played in a national lacrosse jamboree in Amherst, Massachusetts. So I had to fly back here on Sunday. I took JetBlue, had the best connection from Boston to San Jose. Now, the guy that was sitting beside us in the airplane, the young guy, he might have been studying to be a pilot. He might really want to be a pilot. He might have loved the Idea of going up into the front of that jet, taking over the controls, flying us home, he might have felt completely capable of doing that. Particularly with the automated navigation systems these days. How do you think I felt about that? No, I don't want him flying that airplane. I don't want anybody to fly the airplane. I want JetBlue to have a process, a centralized quality control process to choose good pilots that have a lot of experience. Okay? And I want that person to fly the airplane with a co pilot or co captain who's also trained and experienced. So that's an example of. Now, on the other hand, JetBlue is interesting because they did the best job of decentralizing the ticketing and the seat assignment thing. Like, I was annoyed. You can't go up to the counter and ask for a change to your seating assignment. In JetBlue, they're like, we don't do that. Go back to the kiosk. We don't change seating assignments up here. You're like, what? You're kidding me. But they did that to push the responsibility back out to us, the user. So very decentralized, very sophisticated networking approaches. Notice the JetBlue pilots don't have to carry these huge briefcases with tons of documents in them. They carry laptops that plug into the jet, that record the whole flight, the whole flight experience. They're not only going back to central office, they get to keep their own copies. Okay? Very sophisticated network. Anyway, so that's one example. So, you know, we can't say that centralization is better or decentralization. We can say the trend is towards decentralization. Let's give another example of where we want some centralization. Banking. Okay? I mean, look at ebay as an example. In the ebay marketplace, how do you decide who gets to be a buyer and seller? Well, anybody can register to be a buyer or seller, right? You don't even have to put your address in to be a buyer or seller. Okay. But on PayPal, do you have to register who you are and where you live? Oh, yeah. And your Social Security number and your exact bank accounts. Why? Because we're moving money. Okay. And it's not enough to say, well, Rob thinks Kurt's really honest or Mike's really honest. That's not enough in the day to day financial transactions. So you're going to have some centralized control. And you see all these big mega mergers because the banks have to have all this infrastructure overhead to run a bank. And you can cut out half of that infrastructure. You don't have to replicate it when you merge them together. So you get these economies of scale by bringing them together. But at the same time, you've got Prosper.com, coming out that's creating a direct peer to peer market for loans and mortgages. And you got Kiva Kiva coming. Kiva.org, a great game changing play in microlending that allows you to go look at the profile of Jose in Guatemala selling rugs or Kibe in Nairobi running a shoe repair shop. You get to read the profile of a poor person and then choose whether you want to fund them or not and participate that in funding. And you give a zero interest loan on your credit card. And they went from like a million loans last year, which was their first year of doing loans. They'll do maybe, well, lots. A huge multiple of that this year. They're exploding. It's a game changing experience in microlending anyway, so there's not one answer, and it depends, but I'd say in general, the more decentralized we go, the more power we get out of the people engaged in the activity. But we have to have enough structure to make it work and make it a way sustainable system. And I think. And finding that balance is what we call the sweet spot. Next question. Are there any other questions? Yes. Excuse me. Oh, okay, go ahead.

I'll let somebody else go first.

All right.

Yeah. The question, Is it true that centralization gives you more payoff in the case where you're really in a static situation, you're trying to optimize, you're trying to reduce your, you're trying to merge to reduce your overall overhead, whereas decentralization would be more advantageous where you're in an uncertain situation, you've just been chopped in two, and therefore you need to be able to be more flexible. And does that mean that you really need the ability to move back and forth as the situation evolves?

Sure. So the question is, if you're centralized, does that mean in general you should go through more mergers and acquisitions or putting a bigger thing together so you can whack it out those centralized costs. But what, and what does that mean for centralization? Is there a corollary there in terms of maybe MA doesn't make sense? Is that a fair restatement? Okay, good question. Interesting first response. I mean, on the centralized organization, merger side, clearly the more centralized the entity, if that infrastructure is replicated, clearly you're going to have a cost savings. If you can eliminate that duplication, that's like a no Brainer that's going to keep driving M and A until the burden of overhead, of centralization, which is maybe a lack of touch and feel with local communities or whatever it is, outweighs it. And that's again where those organizations have got to find balance too. I was just in a brainstorming exercise with a bunch of CEOs in Amsterdam and someone had rolled up five different major insurance companies and we were talking about this. So given the model of the book, what should their strategy be? And we talked about a bifurcated strategy because in the back end there should be individual winners that are best of category in the different centralized processing areas. But when you merge entities, you've got to distribute that success and failure somehow. Right? You can't just. You don't want one group to completely lose out, or you may not want to in terms of morale. But we talked about the customer interface and they said, you know, one of the big discussions we have is do we want to have one brand out there that's like this huge thing? And they decided no because it's less personal and people have relationship with their neighbor, you know, who's been selling them health insurance for 20 years for all state or whatever the equivalent would be in Holland. And they don't want to change that. In fact, in the downtown area there's an Allstate office and there's a, you know, a geico, whatever office they're at. So we don't want to change that. So in their case, it was a bifurcated. I think it's an interesting general question on if you've got a couple of decentralized organizations, where would you still want to bring them together? And then the question would probably be, are there some great synergies out of whatever intellectual capital they've created or what business momentum or market position they have that can come together as a result of the union of that network? And then also, of course, in decentralization, what you've also got is people will break out and do their own thing. I mean, as the first question referred to the starfish ripping off its own leg. And like I say, people love Craigslist, right? Right. But they love it as long as it gives them this free marketplace to change things. And if they turn on the ticker to charge you money and then Sally's list comes out and it's just as good. You might go to Sally's list just like they killed Napster with lawsuits. Well, then everybody did move to Kazaa because the values the community had Was we want to swap music for free. We don't think we should have to pay 15 or 20 bucks a CD. And I'm not going to get into the moral judgment of whether that's legal or not. Clearly it's illegal, but whether it's moral or immoral others can go and debate over. But the community had that value that we want to swap music for free. The network provider was gone, that protocol was gone. The next protocol came right in because the energy was there, the network was there, and it wanted a way to fulfill its transactions with its value system. And that's why these decentralized networks are so powerful and disruptive, just like we Google's disruptive, powerful entity in the advertising and the online advertising and search world. I mean it's an incredible game changing thing that's happened through the algorithms and through the full set of products and services. Yes.

In the course of your research, did you come across any examples where an organization that started initially as one, like decentralized, migrated and became centralized or vice versa?

We've mostly studied examples of decentralization, but actually when I think about it, there's some interesting examples of centralization we looked at. OK, let's look in the 1970s. The big craze in the stock market in the 70s was vertical integration and the conglomerates. So if you're in the steel business, be in the mining business and then go into the steel part manufacturing. So there was a lot of verticalization that happened in the 70s and there was a lot of conglomerates that came together and they would just buy all these companies and they always had the same argument. We're going to have these synergies. We're going to get the synergies working between our products. Most of those conglomerates failed. Let's look at one that worked. Ge. GE took a different strategy. It bought all these different pieces, it rolled them up together. But then Jack Welch came up with the following set of protocols. He said, well, we're going to run this thing as completely separate business units. And you, the CEO of this business unit, you're responsible to be number one or number two in your market and run it profitably. And then ethically follow these business rules. Over here, you can do whatever you want. And guess what, you don't have to do any favors to other GE divisions. You don't have to sell to them at a different price. You don't have to sell to them versus someone else. You're independent, you're on your own. If you want to partner with Them, fine. But you can complete compete with them. You can compete with another division, okay? Do what you want to do. You have to succeed. You got to be number one or number two or we'll shut you down or sell you. And you got to be profitable. Okay. It worked brilliantly. It worked brilliantly. And GE is an amazing organization. I was in a flight in Europe once, it was like Munich to London or something. And this young guy sitting beside me, 28 years old, I'm like, what are you doing? He goes, I'm president of a steel mill. And I think it was a Czech Republic. I'm like, what? He goes, yeah, I work for ge. I'm like, how did you become president of a steel mill in Czechoslovakia? He goes, well, we bought it really cheap and I'd been trained on how to run business units back in the States and I worked in a metals business. So they sent me over here and I'm president of the company. I'm like, really? He goes, yeah. And he goes, I just got back from China. I'm developing new buyers over there. We figured out from this little mill we can produce some different specialty steels that aren't very common. And we're getting those together because there's demand for these in China to manufacture these new part. And I was like, you're kidding me. He's like doing that. And he's like, yeah, I'm trying to buy new technology and new smelters because we're going to start doing this other new. And I'm like, this is amazing. This guy's 28 years old. He is completely running this steel business in the Czech Republic. And you know, he's going to turn, I can tell this guy is going to turn this around. And GE has the management system to develop those leaders to run that decentralized business model. Now, given that model, what did Jeffrey in Milt change? Right? Because Welch is famous. And Welch did a great job leading George. He's gone. And Jeffrey Anmelt came in, I think like year 2000, 2001. Now what did Endmelt do to the code? He added one new rule. Got to be number one or two in your market. You got to be profitable and you got to be in a high growth business. If you're not in a high growth business, we're going to sell you. Because we're in too many slow growth businesses and it's weighing down our valuation as a company. So we got to get in the high growth businesses and that's where we want to put our Capital, it's working brilliantly. Took a long time. GE got hit, as a lot of companies did in 2000. Their stock was way up, it went way down. Business became tough, was growing slow. They started selling off divisions like Plastics and others that they didn't view as high growth. Now their growth is kicking back in. So he added a third new key protocol to the so that was a huge centralized organization that uses very decentralized management practices. In a flock of organizations that didn't. That trust tried to drive synergies through a strategic planning department that failed. So that's an example of kind of a big centralized player decentralizing. And there's a lot of good examples of that.

Were there any examples of the reverse.

Of a decentralized organization becoming more centralized eBay with the acquisition of PayPal in some sense. So look at ebay. And the whole model was because when Pierre started the marketplace, it was free, right? He created the ebay marketplace every as a free service for his wife and others that wanted to trade PEZ dispensers and other things. And it just started to explode. And six months later, whenever his ISP called him, they're like, what are you doing? You have insane amounts of traffic. You have got to pay us big bucks. And he was like, hang on, I'm offering this free service. And now I got to pay a lot for Internet access. I'm going to make people Pay. I think 25 cents a listing became a business. Then when people started to rip each other off and complain, hey, that guy sold me a broken PC. He said it worked, but that HPPC was broken. What am I going to do? And Pierre, fix it. Pierre, make sure I don't get ripped off. Get the scoundrels out of this community. And Pierre, not being a control freak guy, but being a very brilliant catalyst, very high emotional intelligence, said, the knowledge is in the community. Let's let people rate each other. If people just start rating each other with user ratings on the transactions, well, the honest sellers will percolate up and bad sellers will get cleaned out of the system. So very decentralized in general. Then they bought PayPal again. We talked about PayPal, right? You got to have. I think PayPal has over 1,000 employees that are just doing fraud detection every single day, or 1500. It's a large number. So they had to have a different culture within the organization. Now, I don't know if it centralized the whole thing, but clearly that acquisition in that group operates with a more centralized set of controls and behaviors than the overall organization, other decentralized organizations becoming more centralized. There's other stories out there. Yeah, I mean, even Grameen bank, in some sense the kind of the pioneer creator of the microlending industry has become so huge. It's a major financial institution now in Bangladesh. So it's centralized in some sense, even though down at the coal face it's pretty decentralized. So growth can lead you to have to have control, control structures. And again, they have that constraint in the environment of being a finance company that needs to have those controls. Yes. Okay.

So your third strategy for a centralized organization to compete with a decentralized organization was to decentralize yourself? Yes, and I'm having trouble coming up with examples of where this has worked or understanding competitive dynamics between two decentralized organizations trying to wipe each other out and how that could ever happen. Are there examples that come to mind where this strategy has been successful?

You could talk about the third strategy, decentralizing yourself. Sure, yeah, there's a lot of examples. I mean, one is cloud. I mean, the US Military is having much more success now where they're moving towards special operations when they're fighting this decentralized opponent. There's actually a country in the Middle East, a major country, that got a hold of our PowerPoint presentation from Stanford in 2004 because someone from that country was in the room and asked if they could take it back to the President, the head of their country. They did. Based on our presentation, they created their own decentralized campaign counterterrorism network inside their country and found that it was somewhere between 10 and 100 times more effective with a tiny amount of the resources. So imagine instead of having the western experts or other security experts come in and try to help you figure out where your akitis cells are or whatever, instead you create community based organizations within the slums and within the local communities to deal with the issues. That's a concrete example, kind of in military and police force and intelligence. There's a lot of examples in business. I had very interesting conversations with McDonald's. McDonald's is a traditional organization that took this product set that people seem to really enjoy hamburgers and fries and Cokes or sodas, and rolled that around the world with a very American model, American image, culture, model management systems, everything. They decided not so many years ago to really go localize. You know, they hired a guy in Sweden who was a superstar who ended up opening up 200 new McDonald's, but he customized them to local circumstances and they brought him back as an SVP and made him an svp. Brilliant guy. And they hired other people that said, you know what the problem with us is? We're sending people out of, you know, our area here, all around the world, from the Midwestern US to go run China, to go run Japan. They don't speak the language. They don't speak the culture. They're great managers, right? They wouldn't have gotten way up the system, but they don't know the cultures. Let's go change the whole model and only hire locals. I mean, we have millions of employees. They have 1.3 million employees. We clearly have management talent out there. Now all their key international regions are run by locals. Guess what's happening? Business is flourishing. They have different products on the menu. They have a different look in the new architecture and design of the stores. They may keep the golden arches, but it may look more like a pagoda in China than It looks like McDonald's down the street here off Highway 101. So they successfully began a really decentralized organization. There's, I think, a lot of very successful examples of organizations decentralizing themselves. And again, I think that what this organization is accomplishing and doing on the innovation side is a great example that a lot of other organizations could follow. Although you've got to have the true catalyst leading that transition to make it real. Thank you. Anybody? That's the last question. Right? Thank you so much.