 Tuesday lunches, we're very happy to have with us today Andrew McAfee, who is on a faculty at Harvard Business School. He's actually doing a year visiting at Sloan School as well. He's the coiner of the phrase enterprise 2.0, but more important he is somebody who has over the past couple of years pulled the idea apart and watched how business has been adopting or failing to adopt or not web 2.0 practices and technology fighting for the for what's for what's good in in the enterprise. He's also a Twitterer and a blogger at the faculty blog at the business school. So, book, book, and apparently he has a book coming out. Why do you think I'm here? Apparently there's a book that perhaps on enterprise 2.0 that could possibly be titled enterprise 2.0 probably that could be coming out from a Harvard related publishing house. Yep, sometime this year sometime this year. So, before Andrew begins we're going to first of all have two announcements. The first is that this is all public. It looks like a closed room. It's not. We're on the internet in various ways. So, if there's something intensely embarrassing and personal that you want to say, now is not the time to do it. Just go to Facebook and put it there. So, this is going on a permanent record. And the second thing is that although I don't think we've ever actually announced this before, there is a back channel here which is IRC.freenode.net and the channel's name is Berkman and it's where we make fun of the speaker. It's where we enlarge our understanding and ask ourselves questions and then occasionally snark at the... And I see we're just about out of time. Thank you. So, if you had your laptop open you might want to feel free to join the back channel, which is a very civilized back channel. So, those are the two announcements. Now, before we get to hear you, as we do every time, we're going to go around and if you could just say who you are and what your affiliation is in a single phrase, I will model this for you. I'm David Weinberger. I'm a fellow at the Center. So, if we start here and we'll go around and get to everybody. My name's Anand after that. I'm a night research fellow at NYC. Olivia Jenning. Jonathan Zipfett, longest serving Berkeley term. I'm Matt Sanchez, Harvard Law School student and a volunteer at the Citizen Media Law Project. Thank you very much. I'm a graduate student, also intern with the Center of Stop Backward. Amar Asher, Brooklyn South. Donna DeGeneres, professor at UNAB. Dan Yamcha, new intern at the Berkman Center. Gene Kuhl, fellow at Berkman Center. Dan Lee, director of digital content at Harvard University Press. Erhard Grape, grad of Cambridge, study these kind of subjects. Anthony Lux, law school of communications. Dr. Victor, visitor from the Oxford International Institute. Azure Collier, public relations at Fitchburg State College. And I'm a master's candidate studying social media and generation X generation Y. Alex Howard, associate editor at search compliance. Doc Searls, I'm a fellow at the Center. Jason Kaufman, I'm a fellow at the Center. Dave Larshall, Brooklyn South. John Moore, showmark research, IT analyst firm. Kevin Bracey, Berkman staff. David Hornick, venture capitalist and Berkman groupie. Charlie Nesson, Berkman faculty. Maria Cristina Caballero, former name and fellow at Alameda Kennedy School and journalist. Carolina Hossini, Berkman fellow. Aaron Shaw, Berkman fellow. Adrian Gropper, med commons, health and genetics. Laura Fitton, pistachio consulting, social media. And Andrew McAfee, and I'm Andy McAfee. And for reasons that I don't fully understand, my professional interests have always been around how very boring companies and very boring industries use technology. In other words, you have to study how technology gets produced to understand how it gets consumed. But I'm always most interested in how dog food manufacturers try to take advantage of the great stuff that comes out of the high-tech industries and use it to make more better dog food faster. So what I want to talk about today is the efforts that I've observed inside a lot of organizations to understand and harness these weird new tools that cropped up on the web, get subsumed under this label Web 2.0. So as David said, I launched this phrase Enterprise 2.0 to talk about how mainstream companies are trying to get their arms around this phenomenon. And I want to start just by offering my standard definition of Enterprise 2.0. You'll notice a phrase in the middle there, an emergent social software platform. To unpack that just a little bit, social, I don't think we need to unpack anything maybe except emergent there, which to me means that all of the environments that I'm interested in these days have a couple interesting properties. They start out as very free-form environments. They're usually highly egalitarian. There's not much baked-in workflow. Their decision rights aren't allocated. People don't have pre-assigned roles and identities. But these things tend to emerge over time. And all the tools that I've been looking at have this wonderful property where over time patterns and structure become very apparent. These environments are fairly fine-grained. They're easy to navigate, easy to consume, but they don't start off that way. And this is in pretty sharp contrast to most of the enterprise technologies that I'm familiar with and that a lot of companies are familiar with where you impose structure like crazy with your ERP or your supply chain system. These, to my eyes, are just about exactly the opposite. And we could all rattle off. I'm sorry. Thank you for slowing me down. An ERP system is this massive monolithic piece of software that you install in your big, geographically dispersed company to do payroll and accounting and human resources and procurement and manufacturing and sales and distribution and to just run your business for you. They came online in about the mid-90s. They have taken over corporate technology. And as I say, what they do is impose structure on organizations. Enterprise resource. Enterprise resource. Just layer-in business processes give you very, very little discretion as you're executing them. And they tend to pre-define workflows, decision rights, identities, roles, interdependencies. They just specify all this stuff in advance and then you're expected to go execute on it. All the tools that most of us are familiar with, like blogs and wikis and Twitter and Facebook, they try hard not to do any of that. The huge shift that I've noticed is that these 2.0 tools make a very sincere effort to get out of the way of the users up front. But then they build in these interesting mechanisms to let structure emerge over time. So we don't wind up with continued chaos. There was a wonderful quote in the early days of the internet that the net is the world's largest library. The problem is all the books are on the floor. In other words, wonderful repository of stuff, but you couldn't really get your arms around it or find what you were looking for. We don't feel that way anymore, and to my mind, that's the whole 2.0 phenomenon, how we get emergent structure when we're not trying to bake it in in advance. So when I talk about this stuff to management teams, I sense a level of sincere interest because they've seen Wikipedia, maybe they've looked something up. Their kids have Facebook accounts. They keep on hearing that their company is here, so they're kind of interested in what's going on. But I typically get two very, very good questions back when I talk to management teams about this phenomenon. The first one is how do I understand what these tools are and how do I differentiate among them? I hear about Facebook. I hear about Wikipedia. Are they the same thing? Are we talking about the same thing, or are there important differences in these tools? And the second one is why do I care? I understand this is cool. I understand my teenagers like it. As a pragmatic business person, I'm not in the business of building communities or helping people socialize better. I want to make dog food. Tell me how this stuff helps me make dog food better. I think those are wonderful questions. To launch our discussion today, what I want to propose and put in front of you is the best answer that I've come up with so far to help business people understand what's going on and why it might be valuable to them. And to do that, I want to propose a very, very simple model or a view of what the world of work looks like for a prototypical knowledge worker inside a big, geographically distributed, multifunctional, multi-business unit company. Pick your favorite huge company, and I want to draw a picture of what the world of work looks like. At the heart here for this prototypical knowledge worker is the relatively small group of people with whom she has strong ties. And these are exactly what they sound like. It's a very intuitive concept. These are your close colleagues. You've worked with them for a while. You go to lunch with them. You might socialize with them. You have strong ties by virtue of a long period of fairly intense interactions with these people. Outside that core, I'd say that for almost every knowledge worker, there's a larger group of people with whom she has weak ties. And, again, this is a very intuitive concept. These are folk that you know somewhere between the acquaintance and the friend level. You've worked with them on a project a while back. You've interacted with them sporadically. You know them at least well enough that the classic test here is that if you ran into them in a bar, would you be comfortable having a drink with them? There are a lot of people like that in the world, but a larger group than we have strong ties with, I'd say inside a big organization, there's an even smaller, a larger group of people with whom our focal knowledge worker is not tied at all yet, but should be in some sense or another. In other words, there's a large group of colleagues out there who would be useful to our focal knowledge worker if only she knew about them. They could help her not have to reinvent the wheel. They could point her in the right direction. They could introduce her to somebody else. There's a world of helpful people out there, but at this stage of the model, they're just potential ties. She doesn't know about them, so she can't convert them and exploit them as either a strong tie or a weak tie. I want to stop here and make two points about the bullseye that we've drawn so far. The first one is that we tend to spend a lot more time in organizations. I guess I want to make three points. The first one is we tend to spend a lot of time in organizations strengthening ties that are already strong. So we work on how to design offices for good flow. We do offsites. We do a lot of work to even more strongly tie our strongly tied colleagues. The second point, though, is that there's a huge amount of work in sociology and in management scholarship that points out how important those latter two ties, the latter two rings of the bullseye are. There's a landmark paper written in sociology in 1973 called The Strength of Weak Ties. Don't need to really walk you through the argument, do I, or the point of the paper? We're written by a guy named Mark Granaveta who said, look, we spend a lot of time on strong ties, but actually your weak ties might even be more valuable to you as a person for the simple reason that you tend to know a lot of the stuff that your strong ties know by virtue of your deep interactions with them. They're just less likely to be sources of novel information and bridges to other social networks than your weak ties are. And the initial research on weak ties highlighted that most of us meet our partner through a weak tie, not a strong tie. Most of us get our next job via a weak tie, not a strong tie. Granaveta wasn't an organizational sociologist, but a bunch of people since then have applied his work inside organizations and the conclusion holds up like crazy. A colleague of mine named Morton Hanson at the B School did a lovely study where he looked at work groups inside a big organization tasked with an innovative project and he measured their number of strong and weak ties per group. And it turns out one of the best correlates for how good a job they did, because he was able to grade the results of their work. One of the best correlates of how good a job they did was the number of weak ties that they had. So the strength of weak ties conclusion holds up over and over again. At that third level, there's another massive body of literature about these weird human rolodexes who we all know who seem to add value by making introductions between people, serving as a bridge, a connector, a broker between networks. Is that just a massive amount of very good scholarship here that highlights the different ways that these people add value? And do they do it to increase their social capital? Do they do it for altruistic reasons? These debates rage back and forth, but no one doubts how valuable it is to have someone who can convert a potential tie into an actual one for you. The last point I want to make about the picture so far is that prior to the 2.0 world, or classically inside organizations, we have had lousy technologies. Pretty much at every ring of the bullseye, but particularly at those outer two rings, I usually ask a lot of my audience, especially older ones, because younger ones tend to answer this question very differently, I tend to ask them, how do you keep on top of your network of weak ties? And there's usually this great cricket chirping in the room. The best answer I ever got was someone who said, oh, you mean the Christmas letter. How many of us still get letters from a weird uncle who gives you the yearly rundown on what, and Timmy's soccer practice and the operation that they had in November and all that? And we kind of read those and throw them away, but they're kind of a mechanism to stay on top of what a weak tie has been up to. And I think that historically, and especially inside organizations, we have had very few other mechanisms to stay on top of our weak tie network. At that third ring of the bullseye, corporate directories and corporate white pages try to highlight, flag people's expertise so that their colleagues can contact them. And I've talked with a few of the largest professional services organizations in the world, consulting companies. When you talk at the partner level and you say, hey, how do your people find each other so you don't have to reinvent all the work for every engagement? They say, oh, yeah, we've got great corporate white pages, all the areas of expertise up there. We feed in the engagements they've worked on in industry and all that, and it works like a charm. And I asked my students who typically worked in these kinds of companies for a couple of years before coming to business school, hey, how did you actually find someone to help you with whatever engagement you were on? Did you use the white pages? And they all look at me like I'm insane and say, of course not, we would just send out blast emails or IMs to all of our fellow RAs and start trying to figure out who had worked on this before. So they were using fairly primitive, fairly thin means to try to find potential ties to convert into weak or strong ones. So at those outer two rings of the bullseye, historically, corporate technologies at least have been doing a fairly lousy job of helping us there, even though these are critically important activities. And then finally, we shouldn't kid ourselves for any of our focal knowledge workers. There's going to be a large population of people out there who aren't going to be useful even if she knows about it. Just not going to have a lot of professional overlap. Think of it as the strangers inside an organization. And I used to think that even with the wonderful world of 2.0 technologies, there was nothing we could do at that outermost ring of the bullseye. What technology can help a bunch of strangers come together and generate good output? And I'll show you in a little while just how powerful the technologies can be even at that last ring of the bullseye for bringing together dispersed groups of strangers who are not going to form any kind of professional bond and helping that group of strangers generate some extremely powerful output here. What I want to do before we throw it open for discussion is redraw this bullseye and then talk about what I think of as the prototypical or the archetypal enterprise 2.0 technology at each ring of the bullseye here. So for me, when I think about a group of strongly tied colleagues, I think about any of these group editing environments that we've set up, Wikis or Google Docs and Spreadsheets or any of these online environments up in the cloud that give us all egalitarian decision rights over modifying the content. When I ask my MBA students now how they do group projects, again I get this, well how do you think we do a grandpa look and they all say, we fire up Google Docs and off we go and I try to help them understand that three or five years ago no one was trying to work that way. I usually ask my corporate audiences how many people do group work by emailing successive copies of documents around to each other, about two-thirds of the hands go up. Then I ask them if they're happy with that method of collaboration and all of the hands go back down. So this just seems like a better way to do version control to keep everyone on the same page and you get more work done more quickly, hopefully. I also do want to point out that I'm not making the case that these technologies are only valuable at this ring of the bullseye. It's absolutely not true. We see from Wikis and Wikipedia that these can be wonderful technologies for bringing basically very disperse groups together. But I place them right at the innermost ring because if you do have a group of strongly tied colleagues these tools, these group editing environments can be very productive for them. I saw this firsthand at a local company called VistaPrint. They do business cards and stationery and things over the web. Very fast-growing company, very dynamic, bringing in tons of engineers and they needed a knowledge repository for the organization. They didn't want an old-fashioned knowledge management system. They didn't want to have binders full of procedures that gathered dust on the shelf. So one of their senior technical people launched a Wiki and his technique for getting it populated was really interesting. He put an initial structure up there and then every time he saw an email flow through the organization that had content relevant to the Wiki he would just say to the sender, would you mind just putting it up there underneath this heading? Don't worry about formatting, don't worry about linking. I'll take care of all that later on. Just start throwing content up there. He found that was actually an effective enough way to get the ball rolling in the right direction and now it is, he says it's absolutely the default first place you go to learn something about the company. Very dynamic, growing all the time. He put a thousand topics up there initially and it's grown basically 12,000 different pages on there. At that second ring of the bullseye this is where our friend's social networking software can come in extremely handy. When I think about my Facebook account, and actually these days, Twitter for some of us at least has replaced Facebook as a way to stay on top of your weak time network. These are really something new under the sun that can be very powerful tools for helping you build and maintain and exploit a very diverse network of people who you're not very strongly tied to. You get I think two flavors of benefit from doing this. The first is that you get information you would not have had otherwise. So when I look at my tweet stream, the people that I follow on Twitter every day, I get links to articles and bits in the news and things that I absolutely would never come across in the course of a normal day. I find it tremendously helpful. The second thing you can do is use it as a launchpad to other social networks. So these tools are wonderful bridges to other networks out there. Again, I see this in Twitter when someone retweets something that came from someone that I didn't know and then I can follow them or I can build up a network that way. I think these tools are great not just for building and maintaining this kind of network, but also for exploiting it. I use both Facebook and Twitter to throw questions out there and a very low overhead way to do it, very low overhead to help somebody out or answer a question as well. So the barriers to helpfulness or colleaguehood have dropped through the floor with these kinds of tools. I saw this in action at a company called Serena, a back and old mainframe software maker now tries to do enterprise mashups, grow by acquisition all over the world, very geographically dispersed workforce, and the leaders of the organization were worried that they were completely losing the idea of a shared company culture or any kind of identity for the corporation. They were turning into a group of freelancers who all picked up paychecks from the same place. To try to deal with that, they encouraged all their employees just to get a Facebook profile up, and the way they got that kickstart was having the first Facebook Friday where they said to all their employees, look, just show up dressed in the clothes of some activity that you enjoy. That could go all kinds of wrong, right? But this actually worked out fairly well. The CEO raced cars, he came in in his racing suit, people in golf clothes showed up, they took pictures of them, they put them up on the Facebook profile. They said one of the first things they noticed was a picture of a colleague up there. You feel the glue a lot more strongly and you're much more likely to reach out to that person or to try to interact with them. After a little while, they found that 90% of the people, this was optional, 90% of their employees had at least put up a profile, 50% of them went there at least a couple times a week, and about a quarter of their employees were junkies with Facebook, spent a lot of their time there. I said to my contact, that's all well and good. What business benefits has this realized you? And he said we're having our user conference next month. We have twice as many people signed up for it as we did for the last years, and we've done absolutely no traditional marketing or outbound marketing for the conference. We just encouraged our employees to talk about it in their Facebook, MySpace, whatever profiles and to answer questions from their professional colleagues. He also said we're really trying to refashion this company so that it's a cool place for younger people to work when prospective job candidates see that all of our employees are out there and participating in the social media. It helps us rebrand ourselves as a workplace as well. So we're getting a lot of our new hires that are coming in via these tools as well. At that third ring of the bullseye and converting potential ties into actual ones, a couple different ways to think about this. When I think of an active internal corporate blogosphere, I think of this as a phenomenal way for people to talk about what they're doing. The great phrase I heard about blogging is narrating your work, and if you put some decent enterprise search on top of that, you've got a very effective way to find out who else in the organization is interested in the same things you are or doing the same things you are. It makes the problem of finding somebody else trivially easy to solve, and it helps you convert potential ties, even in a big organization, into actual ones and then to work with them from there. So the benefits are that, hopefully, your rates of innovation go up, you get better bridges to other networks out there, even in a large organization. I'll show you a couple examples here. This is a screenshot from a real old-fashioned company called IntraWest. They make ski resorts, lodges, in the western part of the U.S. and Canada, and a while back, they basically replaced their 1.0 intranet with a fairly dynamic 2.0-flavored version of it that gives every employee the ability to blog or to do something very similar to blogging. What you see here in four paragraphs and three pictures is one of their employees blogging, essentially, about a new way that he found to lay a radiant heated floor in the lobby of a ski resort. How many of us are deeply interested in radiant heated floors? Wow, surprisingly high percentage. Well, where this gets interesting is that he said, the way I found to do it saves half a million dollars on the baseline cost of $2 million for doing it. So here's what I did. It didn't take that long to describe, and he put three pictures up there to help his colleagues understand what he had done differently. Now, where this gets really interesting to me is down at the bottom of that page, that's a comment to his blog from a colleague. Saying, hey, this is interesting. Can you upload a couple more pictures to the corners, which is evidently where this gets difficult. Now, in a sense, these are two contractors talking about laying radiant heated floors, which is not that exciting to most of us in the room. In another sense, this is a half million dollars in savings transmitting itself throughout an organization. This is the kind of thing that gets my corporate audience is fairly excited whether or not they care about radiant heated floors and ski resorts. I'll give you one other example at this level. The organization where I saw this the most clearly is the United States Intelligence Community, which is the group of 16 agencies charged with intelligence analysis. So this alphabet soup of CIA, FBI, DIA, NSA, GSA, you can go on down the list. If you remember, these were brought under one bureaucratic umbrella not too long ago by either a presidential or an act of Congress. I can't remember which. After 9-11, the 9-11 Commission said the lack of connecting the dots is one of the great failures of 9-11. And one of the solutions was we need to have an umbrella organization over all the intelligence agencies in the country. When they announced that, when the DNI was formed, which is the umbrella organization, a lot of the commenters said this is just a really, really poor idea. What we do not need in the intelligence community is an additional thin layer of bureaucracy. We've got plenty of that already. And they said, what can the DNI do to overcome this extremely entrenched culture of not getting along and not sharing and protecting your bureaucratic turf that's so well ingrained across all these different agencies. These are huge, well-funded, deeply entrenched, high-culture organizations that do not like each other and don't share very well. I was doing interviews at the CIA. I said to my contacts there, come on, is this really true? Do you guys really not like the FBI expecting them to go? No, it's a little bit overblown. Instead they said, we hate those guys. Our organizations are just deeply hostile to each other, essentially. One of the things that DNI did do and could do, though, was mandate some technology that would go across all 16 of these agencies. And the agencies did not have the option of saying, no thanks, we're going to pass on this one. This is actually something a thin new layer of bureaucracy can do that's very, very useful. And they were smart enough to deploy a set of 2.0 tools across the entire intelligence bureaucracy. So there is one wiki environment called Intellipedia that goes across all 16 agencies now. There's an internal blogosphere where if you have the appropriate security clearance, no matter which agency you're in, everyone else's blog's at that level of security clearance. They have a tagging environment, they've built the equivalent of YouTube and Flickr for this, and they've put this nice suite of 2.0 tools layered across all 16 of these agencies. And I went to go talk to them about how they were using it. And in best 2.0 fashion, one of my contacts said, why don't you send me some questions in advance? I'll post them to my blog and we'll see what kind of answers we get back from throughout the community. The first question I asked was, please tell us what benefits you've realized from either using these tools or at least contributing to them or at least consuming them. What I was expecting to hear back was something like the Intellipedia entry on poppy cultivation in Afghanistan is now the single best source on that topic inside the entire community. It's become the go-to place for knowledge on this topic. And instead what I heard back was something a little bit different, which I wasn't expecting, but I thought was super important. They said instead, because of that Intellipedia article on poppy cultivation in Afghanistan, I now know someone who works inside the geospatial agency who's contributed to that article and who clearly knows a lot about the subject, has analyzed all the satellite photographs and has really put some good content up there. This is a person who knows their stuff and I had no possible way on Earth to learn about that person's existence, to understand their work, or to find them and be able to interact with them back in the world prior to this. So what these 2.0 tools have done is not so much giving me access to a pool of great information. That pool of information is a reflection point for me to find other brains inside this huge bureaucracy that I can then pick up the phone and talk with later on. I had no way to find those kinds of brains before. Last thing on the bullseye, this universe of people who really don't have anything to say to each other, they are strangers and are going to continue to be that way. Is there any piece of technology that can help us at that level? When I look at prediction markets, both on the internet and corporate prediction markets, what I see here is a phenomenal technology for harnessing collective intelligence, the wisdom of crowds, whatever buzz phrase you want, but even if these people don't know each other and are never going to become good colleagues, these markets that they harness the wisdom of those very dispersed bodies of people and they really do generate something close to collective intelligence. We see all kinds of evidence of this. Google has, to my knowledge, been running the largest and oldest corporate prediction market inside any company. Every year they get about, every quarter they restart the markets. They get on the order of a thousand traders and maybe somewhere around 10,000 trades. Compared to the scale of Google, this is not a lot. What they've found is that there's a small core of people who are hardcore traders and a larger group of people that changes every quarter who do a few trades because they're interested in one of the markets that gets set up. One of the things that surprises me still is how small that minimum efficient scale is. In other words, you don't need to have thousands upon thousands of traders and millions of trades before you get really good accuracy out of these markets. I'm talking with a vendor of corporate prediction markets about putting this in my MBA class and I said, I'd love to do it, but I don't want the lesson to be that these things don't work because they do seem to work and I'm worried that I don't have enough students to get a good answer out. He said, how many do you have? I said, this semester I'm teaching a class to about 70 MBA students. He said, you're fine. You get a very accurate answer out of these markets. So our prediction market is simply a stock market. Where the securities you're buying and selling and the securities that comprise your portfolio are not related to the value of a company, as is the case with the normal public stock market, they're related to some other future event out there. They really became popular for the first time when the University of Iowa set up a market on the, back then it was the internet, the green screen internet for presidential elections started in 1988. So if you went to a bunch of the different markets in the 2008 presidential election you could buy and sell the Obama stock versus the McCain stock, for example. And you can set these up in a couple different ways. One way to do it is what's the probability that Obama is going to win the election versus McCain is going to win the election. Boy, that security is set up is that you get a dollar for every share you have if Obama wins and you have the Obama security and nothing if he doesn't. It turns into a market where the collective probability, the collective best guess about how likely Obama is to win the election is reflected in the price. So prior to the, right prior to the election the Obama security was trading at about 80 to 85 cents, which meant the market as a whole was really confident that he was going to win. You can set these up a bunch of different ways, you can set them up so that people are buying and selling what they think the final percentage of it's going to be for each candidate, but the vote breakdown. That turns out to be really accurate as well. They did, I saw at least one where they were predicting final electoral vote totals for each of the candidates. This is hard to do, this is extremely hard to do. Again, they were phenomenally accurate and I think if Missouri had gone the other way, remember how raised within Missouri was in 2008, if it had gone for Obama instead of McCain then the market would have been exactly right on it. So very hard things to predict. These prediction markets spit out really good answers. One of the pieces of research done about the Iowa markets compared the accuracy of the market at different points in time to the accuracy of the polls these big nationwide expensive polls Gallup, Reuters, Zogby, Quinnipac all that that we get bombarded with leading up to the election. How accurate are they? It turns out that at every point in time, you know eight months out, six months out, one month out the market is closer to the actual final total. It's more accurate than the big expensive polls are. There are a couple exceptions to that but overall these markets work and when you think about what's going on any of us could open an account in the Iowa market you play with real money but not a huge amount of it. So where this dispersed group completely uncredentialed you don't have to be a political junkie to start buying and selling but this dispersed group of strangers buying and selling these securities leads to more accurate answers than extremely thorough well done professional polls when you look inside companies you see the same thing. These markets tend to spit out more accurate answers about how many widgets are we going to sell next quarter than the forecasting department and to me it brings up the immediate question why do we still have forecasting departments inside corporations they seem to be doing a worse job than the collective wisdom of the employees still is. There is at the end of the quarter. So now I'm done basically with the formal part I'd love to talk about wherever we want to go. Let me pick up on this question. Google does three different categories of question the first one is internal corporate events of interest how many Gmail users will there be by the end of the quarter are we going to buy YouTube or not so salient events inside the corporation the second ring is stuff inside will MSN search the new version of it launch on time or not will Yahoo and Microsoft merge or not and then at the third level they do pure fund markets to entice people to come in and to trade so how much money will Star Wars make and it's happening weekend who's going to win the world cup they set up markets for that each of the traders can buy and sell any basket of those securities and you get rewarded based on what your total portfolio is at the end of the quarter does that answer the question I'm curious about the point you're making here about why is there still in-house corporate forecasting just like you might ask why are there pollsters and is it possible that the prediction markets would work considerably less efficiently if you didn't have the pollsters or the forecasters just as if we all bought index funds nobody would be looking at how corporations are actually working and then nobody would know how to price anything so is all of this like Web 2.0 stuff layered on top of just like bloggers still rely on Germans to actually turn up a lot of news and it's pretty clear that the traders in the Iowa electronic markets are looking at the polls right and seeing if they're correct or not so I think that's an interesting point how important are the previous categories of input to this question in the trading that's going on I would love to do experiments where there's no forecast at all being done my intuition tells me the traders would be just about as good as they are anyway but they're clearly the point about these markets that the traders are somehow taking in all available information and doing some filtering or refining with it to make it better than it was before okay I have a few related questions first why is what's the reason behind this prediction market why is accurate and the second why is if that's accurate why we still use the poll we can't just use the prediction market and replace how they get existing poll firms so two wonderful questions the first one I'm going to turn it over to one of the great economists of the 20th century Frederick Hayek was an Austrian economist part of the Austrian school in the middle of the century and his work is completely amazing to me because in the middle of the century in Europe most of the intellectuals were just enthralled with collectivism and we were all looking toward socialism or communism or some kind of very very very directed economy as the way forward for humanity and the Austrian school in particular Hayek blew the whistle on that and he said you have got it dead flat wrong and he said when you look at a market what we tend to do is look at it as a way for some people to get rich and some people to get poor he said put that aside for a minute what we've really got here is an absolutely brilliant system for aggregating and transmitting information throughout the world and we do that via the price mechanism he said we have to look at it as such a big deal that it would be a claim as one of the great triumphs of the human mind if it was invented de novo somewhere we'd clearly hand out every price possible to this because it kind of grew up over time it gets folded into markets we tend not to pay attention to how big a deal the price mechanism is so that's an extremely quick answer to the question of why do these work so well they work because there's a massive amount of information conveyed in a price and all the traders weigh that against their personal beliefs and preferences and decide if they want to buy or sell more basically the other question is why do we still have pollsters because pollsters like to stay in business and as long as there's someone consuming their products and willing to pay for them they'll continue to generate polls I would imagine that the market share of polls is going to go down over time as we get more experience with how well these things work because as I said even in the 2008 election the prediction markets did a fairly brilliant job here's a market of the price for that winner take all security over time and what you see is one of the classic things that we learn from markets they change as new information becomes available we expect prices to be dynamic so here we see obviously blue as Obama red as the king you see a couple events I believe this is the republican convention when Sarah Palin was introduced and there was a huge amount of enthusiasm for her and then as more information got revealed the prices started to have the other direction so I took this picture at the end of October and there you see the Obama security is trading at about 85% the market was really confident at that point that he was going to win the election one last thing I want to point out people try to manipulate these markets one of the things that we expect to see in any kind of market these attempts tend not to work very well so what I see here is pretty clearly someone trying to drive up the price of the Obama security the ability of one trader to do that and to keep that differentiation or keep that spike up there you see what happens the market sees that and realizes there's money on the table buys that security back up or sells it until the price goes back down to the correct level the collective level there my colleagues at Google have told me that some of the fiendish geniuses there have tried very hard to manipulate the internal corporate prediction markets just to see what happens and even in those relatively thinly traded markets manipulation tends not to work very well people realize there's money on the table and they go pick it up very quickly about the change in from doing say randomized controlled trials to doing using Web 2.0 methods or other methods for doing science can you make that give us an example or say a bit more about that so can you compare the two different kinds of doing science sure a randomized controlled trial is the gold standard yes it's a very structured environment the alternative is to run a wiki sure if you would and and then process the outcome as a meta analysis if you would or even actually primary data okay I think I see where you're going so to make that distinction slightly differently why does it work to set up a prediction market for will this drug work rather than doing very expensive randomized controlled trials about its efficacy my gut tells me that market would work fairly well I doubt the FDA would accept it though because I think of these as in many many cases a compliment instead of a substitute for existing activity when I look for example at corporate planning and forecasting activities they are intensely political processes and when I've looked at them the goal is not the right answer the goal is the answer that everyone can live with and those are not the same number for how many widgets are we going to sell next quarter reason I like these this this tool of the prediction market so much is I think it's organizational truth serum in some ways that organizations really do need the truth in addition to looking at this as a change in sociology of company or certainly change of the technology happily you look at the sociology first you could also look at this as a power issue a sort of political issue so how does this taking the enterprise as a massive power structure with intense embedded interest I've never observed that at all how does this get accepted what are the threats they're seeing does this change that reinforce I think this is a wonderful question maybe the deepest one going forward in the discussions I've taken part in to now they tend to be about how do we get critical mass going how can we convince people to change their ways of working and use these new tools get them over the barrier and I've asked David's question over and over which is have you seen entrenched interest in this organization actively hostile to this stuff and the answer I've heard back over and over is no and I find that amazing I basically think it's because we don't have enough track record with these kinds of things the fact that they are organizational truth serum and in some cases really profoundly upsetting the status quo that's not widely and deeply enough understood yet as it is I think there's going to be another really interesting set of issues and decisions inside companies are we going to support this are we going to work like hell to maintain the status quo the organization where this is played out the most in my experience is again the US intelligence community the debates over Intellipedia and the 2.0 tools have been very vocal and quite sharp in some cases so far the leadership of the DNI has said gang we hear what you're saying our explicit goal and McConnell in one of the statements to Congress said we are going to try to move this culture from need to know which is probably the most deeply entrenched part of analysis culture in the United States our explicit goal is to move that from need to know to responsibility to share could not be a bigger deal for the spy world these tools are a huge in my eyes a huge part of that transition so far inside the community they've lined up the walk with the talk just stay tuned going forward follow up briefly that's a great example in the intelligence community wonderful example but because it's so explicitly hierarchical and the roles are so explicitly defined as in the military straight forward militarization the role of weak ties and the strong ties you don't have people jostling as much for position for power by their position in the amorphous social network one might expect that the resistance would actually be stronger in a more lateral organization more amorphously defined one way I could see that as well and like you say the way to get ahead in the intelligence community has never been to go talk to your buddy in the FBI it's just not how it works let me tell one quick anecdote about the way that the community is changing the way you get ahead classically in the intelligence agencies is by writing the really brilliant cable that goes to your boss's desk and sits there and if it's good enough and timely enough it goes up one level and winds up on the secretary of defense's desk and you get called in for a briefing the information is incredibly stove piped it goes up the channel and there's kind of a decision point about passing it on at every level of the channel that's not too much of a caricature about the way it's been done historically with these 2.0 tools I did a really interesting interview with a young analyst who said you know I walked in the door of this organization and I wasn't a big blogger externally or anything but I looked at the work of analysts what they were asking me to do and I just thought these tools were a flat better answer for doing it so he said what I tried to do is I launched a blog and I just started getting my working hypotheses out there which is again just not what you do in the 1.0 world of analysis you put your fully baked ideas and conclusions on your boss's desk and I started to work in a really really different way and just blog about what I saw going on and what I was thinking at the time and that blog started to become really popular and I'm pretty sure that this is unclassified so none of us are going to get incarcerated for hearing it but he launched a blog called Ayatollahs R Us by title alone became one of the most popular blogs in the community and he started talking about the theocracy in Iran which is an area of coverage and what he saw going on there it became one of the more popular blogs I think in the top secret level of classification and he started to gain a lot of human capital inside the community and even inside his vertical organization by virtue of doing this obviously smart thinking throughout the community as a whole is that guy a huge outlier? Right now I'd say he still is are there going to be more examples or fewer examples going forward? We have to stay tuned So in telepedia we see that's a top down approach and for example the same is happening in USAID with the global development commons and I was a consultant for that but my point is when you have what you hear from managers related to fears of compactors because sometimes this and this blog etc. they are internal but some when you provoke this culture a lot of information becomes public information so how these managers how people that are building these are fostering this culture they are leading with this fear of compactors or it's just a responsibility to share approach that they are trying to use So if I'm hearing you right you're asking me about security concerns about this information going to the wrong people? Yeah I'm not thinking now about the telepedia but these models in general how open or how public how shareable they are and how the managers deal with that maybe I'm changing ideas with my colleague and I put something that's a kind of confidential information of the company Absolutely and it's one of the first concerns I hear when I start talking about the brave new world most management teams immediately highlight these different flavors of risk or threat from too much sharing there are a couple different responses to that one is we have e-mail accounts and USB drives and Xerox machines inside almost every organization that I'm familiar with if someone wants to take sensitive information and walk outside the company with it they have 38 ways to do that right now what these tools do is make a lot of information more universally visible inside the organization and to the extent that knowledge is very very important or valuable it increases the number of people who have the ability to go take it outside the organization it's an absolute real threat again I can't think of an organization with deeper security concerns in the US intelligence community and they've decided that the virtues and the benefits of sharing better outweigh the risk and they said in very clear terms to me look we understand that if too much sharing goes on people are going to die we also understand that because enough sharing didn't happen many people died but we're going to swing in the other direction here adopting also closed models or they are for example Pfizer has a wiki to discuss how open this there's a big range in general I think too many companies close down their environments too much in other words they stay at this level of a strong tie and they say here's a work group that might be a research lab for example they need a collaboration environment so we're going to set one up for the lab what we're going to do is make sure that only lab members have access to that no one else can see it no one else can contribute clearly times when that's appropriate the problem though is that you foreclose the possibility that someone from a very different part of the organization might have a good idea or a valuable contribution to make and I think we underestimate the value of that and the value of that kind of serendipity a colleague of mine at the business school a guy named Karim Lakhani did a really nice piece of work on Innocentive have we heard of this organization Innocentive was a spinoff from Eli Lilly a huge pharma company and what they said is okay most big R&D companies have a bunch of labs and they have problems that they're stuck on there's some flavor assigned to the problem that they're just scratching their heads on they can't find the answer what we're going to do, we're Innocentive we're going to take those problems from these labs we're going to anonymize them so you can't tell that it came from Merck versus Pfizer we're going to post them up on the internet and anyone in the world who wants to can download that problem statement and if they've got the answer after a while they can upload the answer the other thing that we're going to do is handle the IP transfer so that there's no debate about who owns this and we're going to if the problem is actually solved we're going to do the money transfer so those cash rewards from doing this when Kareem took a look at think a hundred plus problems that were posted to Innocentive he found that about a third of them were actually solved and you can look at that number as low I think that's incredibly high given that the lab itself was stunk was almost a prerequisite to posting the problem the best determinant the best predictor he found for whether or not a problem was going to be solved was the diversity of scientific interest among people downloaded it and he's got all these wonderful examples of someone who uploaded a crystallography problem and someone else said no no no no that's an organic chemistry problem you're looking at it all the wrong way so it reminds me of this great quote from Eric Raymond that with enough eyeballs all bugs are shallow and I think what we don't do enough of it certainly inside organizations is letting enough eyeballs look at stuff is that a fair answer um great examples and I think you framed these approaches excellently this is how to start off comment by the way we use this as a template I haven't gotten to the butt yet no drag! next no I want to play skeptic for a moment and I want to ask you know what it seems to what do you have any sense of the extent to which there are get the sense that you're seeing confrontations around these techniques within the firms that you were looking at within the industries that you were looking at is there any way to generalize out to the level of what tools are proving more useful what industries are these growing fastest in what kinds of problems are they addressing most effectively so let me give a couple these are just impressions not backed up by anything systematic at all what kinds of technologies are most likely to be deployed right there at the level of strong ties the wiki for the work group is probably the classic example and I've talked to very big organizations who are proud of the fact that they have set up 6,000 different workspaces all throughout the big globally distributed company that's great in a sense my huge problem though is I guess two of them one that you can't get the other eyeballs looking at stuff and two that you might not even be able to do things like linking and tagging from one of these to the other so if you've got 6,000 mutually inaccessible walled gardens how much, you know, what's the group level benefit here as well and I always think about the web web and the massive benefit that we have is that there is one web there are not 6,000 mutually incompatible webs I see a lot of organizations fragmenting, deliberately fragmenting their internal 2.0 environments I usually encourage them to think long and hard about why they're doing that and what they're foreclosing as a result as far as what kinds of industries not too surprising the tech industries themselves tend to be relatively heavy users of these things their workforces are young their techno files by definition but I heard something really interesting from the senior HR guy at Yahoo I told an anecdote about a different big tech company that I visited where they had their internal blog role because they had set up a pretty thorough internal blogging environment they said, hey, look at our blog role the first post that I saw was from someone fairly low, often some fairly low level technical part of the organization and the title of his blog was why our corporate strategy is all messed up and so I said to my host I assume that guy's been fired and they said absolutely not, this is the kind of debate that we want to have inside this company so I told that anecdote and I think the head of HR for Yahoo said, you have to understand in our industry the human capital is so scarce and so mobile and these people are so used to expressing themselves very freely, any of us who have spent time in geek heavy environments they're not shy about their opinions he said if we repressed them if they felt like we were repressing them they would just move down the road and go work for somebody else so we have had to learn to accommodate this level of pointy elbow debate inside the company when I tell that story to more mainstream audiences and I say how many of you would be comfortable with that absolutely you know fewer hands go up in the road is there another part to the question and I don't want to give the impression at all that Enterprise 2.0 is a title that is in the process of sweeping over corporate America that we're 50% there or anything, it's just not the case I think we're in the early days there is a huge huge range in receptivity to these ideas among the audiences that I've worked with managers looking at his team and saying you better find a way to get us some internal Facebook because I believe this idea of weak ties to people listening to every piece of advice and evidence that I can give them and saying that's interesting it's just not for us, see you later we've got other things to do but to my eyes that variance is really interesting because that's going to increase differentiation in the marketplace please I live in the people's campus and one of my contacts is the driving incident where that kind of behavior added to the real world up to the calendar level really? in the driving incident because it started online and migrated to the real world now it's in the elbow of a behavior and in California there was a traffic incident and one of the things I try to do when I talk to corporate audiences is collect horror stories I say tell me the worst things that have happened from your experiments or your work so far with these technologies and I try hard to listen to them my collection of horror stories is just about empty over and over again I've heard people telling me that people are incredibly respectful or careful and they do not use it as a way to post hate speech, harass their co-worker put pornography up do any of these things that tend to frighten people in advance the best the same absolutely and I actually think that because these digital environments are very visible and inherently group level there's some behavior moderation that goes on and if you look on the internet all of these geek screaming matches break out all over the place one of the things to keep in mind is that on the net the default is anonymity or pseudonymity and inside the organization the default is identity so my comments are typically tied back to me and if I'm misbehaving my boss comes, the HR director comes or my colleagues come and say look would you knock it off what we want to do with this environment so the community policing tends to work well there's definitely a lot of modeling going on but again I've asked with all sincerity tell me the worst that's happened and I just walk away hungry from that discussion one of my colleagues said you have to understand by this point in our company's history people know how to behave and how not to behave they know what's going to get them fired they know what violates policy maybe we have to worry about the 22 year olds four years ago entering organizations we really don't have to as much anymore I try hard not to be a Pollyanna about these issues but again my collection of horror stories is very very thin so we've had web 2.0 their BDT screens and their flat screens that kind of corresponds to 2.0 I have the sense that it's really very early that and so I'm wondering if you agree with that and where you think it might go is it more versions, is it more fanning out mobile versions, we have static we have live, we have other things I do agree that inside the enterprise it's very early days and this 2.0 label is a perspective for the most part and hopeful for the most part I contrast that to the web where I think the 2.0 is a description of what's actually has taken place or is taking place right now if you go to Alexa look at the most popular 10 websites in the world right now I think 5 or 6 of them are clearly 2.0 websites they're community based, they're emergent they're all these properties so that versioning on the web is completely appropriate I think early days, I completely agree really early days inside the enterprise now as you can imagine there's already the drumbeat of web 3.0 a lot of people combine that with the semantic web and burner Lee's vision about where this is headed I find that really really speculative and to my eyes somewhat unlikely as well his vision like you know of machines doing a lot more talking and interacting with each other with a lot less human interaction and to me just the beauty of the 2.0 phenomenon is that it is about the people and not getting the people out of the way and we like that the 3.0 is another one of these I'll caricaturize it a little bit another one of these geek utopias where the computers are doing all the fun stuff and the people are relegated somehow I don't like that vision to me the huge insight behind is let's put people back front and center here and not try to get them out of the way please 7 years ago in Korea I was selling social networking software solutions to smaller companies and the companies used they were actively used in 2 or 3 years but after 2 or 3 years the employees really didn't want to use them because it was in addition to their work and so I'm just wondering if this is like a short term and I think in a lot of cases it will be one of my friends is a guy named Michael Eidnopoulos who works at social tech which is an enterprise 2.0 vendor and he makes this really smart distinction between work you do that's in the flow of your job and to the extent that participation of these things is above the flow I think it's going to be a much smaller percentage of the employee base and like you saw it's very likely to fade over time so I did a series of blog posts a while back that led to some fairly pointy elbow debate where I advocated that if you're serious that this is an organization measure it rate people on it make it part of their job description and if you want to have a dynamic environment put incentives in place and measurement in place for it some people out of respect very very much have taught me a lot said Andy that's 180 degrees the wrong way to go and if you want to kill this thing measure it compensate on it reward it do all these boring corporate top down heavy handed things but the reason I articulated my point of view is I share that concern if this is peripheral to the life of the organization it's much more likely to fade away over time and I keep on hearing leaders of companies tell me you know you can hear the corporate boilerplate people are our most important resource you want to help people self-actuals and I say look you've got a tool set that does exactly that you know that it's incredibly popular out there on the web put your money where your mouth is and the range of responses again is huge there please put it on your forecasting prediction hat I'm going to set up a market right here right and I'm looking at this enterprise 2.0 okay you said we're very early in the early adoption stage right now with some leading innovators bringing that into the enterprise and using these tools okay when will we see mainstream adoption or when will we start to see in the market a divergence between those that are using these tools effectively and actually winning in the market versus those that are laggers that are not winning and therefore all of a sudden we've got to get religion and get on this that is a fantastic question some research that I've recently done the divergence exactly the divergence that you're talking about is clearly taking place in the economy so a colleague of mine, a guy named Eric Ringolson at MIT and I have done a bunch of work to try to understand not whether IT is influencing productivity because it pretty clearly is by now but is it actually separating winners from losers over time or is it irrelevant for competition or is it the great competitive level so we grabbed a bunch of data from throughout the economy it turns out that the more IT an industry consumes in other words the more IT investment there is the more winners have been differentiated from losers since about the mid 90s and competition in those industries has become more turbulent more winner take all and there's been an increasing divergence in performance in ROA you know gross profit margin market cap for dollar all the things that an investor or manager cares very deeply about so the evidence is growing that either something else altogether is going on or IT is increasing the nastiness of competition basically right now we can't tease apart what kind of IT is driving that difference and because it's so early days for all of this stuff this stuff can't be making much of a difference yet my forecast my prediction is that that divergence in performance and the nastiness of competition is going to be heightened because of exactly these technologies life cycles everything else we start looking at innovation with the weak ties and the potential ties exactly and because of that I would expect the competition in the industry to have exactly those characteristics advantages a lot more transient the good idea gets replaced by the next good idea very quickly the analogy that we like to draw is you look at something like the cell phone market and it's an incredibly turbulent market remember when Motorola was red hot then Nokia was red hot now Apple is red hot research and motion are they rising or falling that's because the next good product idea just determines success or failure in that market you can rise and fall very very quickly we make the same analogy to business processes inside other industries and say it's the way you're going to take care of your customer or fine tune your value chain that's going to help you rise up but someone else can have the next good process idea and they'll take over your leadership position so we clearly see that going on I think it's only going to speed up can I jump in for a second because in his Harvard Law School after all so I want to direct a question to David Horvick who knows a little something about investment and the like whether you agree with Andrew's researchers showing him that this stuff is me too I think it's interesting I look back in the early 2000s and there were a number of companies that were emerging trying to bring social relationships into enterprises in more specific ways so I don't know if you recall Spoke I forget whether it was called Invisible World or Visible World there were some of the early guys that were trying to extract kind of social data and drive it in and ultimately the winner was LinkedIn which was the most which was most clearly outward facing if anything it's viewed as a way to get out of your organization bring people into your organization not to look internally so I just think that's it's just an interesting question of is it that mainstream social products are going to be the things that have to be brought into the enterprise as opposed to things that are specifically enterprise driven but that doesn't answer this question of do we think that that investment in technology is going to continue to grow and continue to grow companies I'll just give one example which is I'm an investor in a company that's doing price optimization for banks now you would think this is just a catastrophic time to be selling multi-million enterprise solutions to banks and in many ways it is because just when you think you've sold one the bank either goes under emerges with someone else on the other hand when these enterprises look at the possibility of being able to optimize versus optimizing I think that they recognize that efficiency as you say efficiency levers are so monstrous using technology that it's almost impossible not to utilize those that are available and so the only question is how long does it take for something to be recognized as actually working the second it is working it has to become part of the mainstream and those that adopt it and so I think what you're talking about is an interesting question which is when will these social media experiences become recognized as mainstream I'm an investor in the sixth part in the blogging space and obviously blogging has had this evolution relatively quickly in the grand scheme of things but now is Twitter supplanting mainstream blogging is social networking so one of the problems I see is that it evolves sufficiently quickly that just when you think you can get your arms around the lever that is the winner Twitter emerges and you go oh that crap you're so old school Wikis are so 2004 and a lot of the management teams that I talked with want the pace of innovation around technology to slow down because they're tired of not having blogs be the cool thing anymore and now it's not even Facebook it's Twitter and they're like would you please just stop the innovating the answer is no that's never going to happen either can we do one more is there one more grab a sec so you know I've done a lot of work on enterprise Twitter and I'm really very pony blind or centric but that's going to be something that can spread and catch on because it's so lightweight it's so simple pick my arguments apart I'm with you on this one I think that it does a couple things really really well I kind of put it at that somewhere between the second and the third ring of my bullseye and it's a great tool for converting potential ties into actual ones and you've helped me see this in a lot of my own work on Twitter the other thing that I really like about Twitter they've got two technical aspects right one is the asymmetry so on Facebook I find my Facebook page clogged up by all these people that I was kind of polite to and I became Facebook friends and all of a sudden all their updates show up with people that I really care about so it kind of clogs up my stream with Twitter you have the choice about whether or not you want to follow someone even after they follow you so you can kind of shape your network a lot better the other thing is like you say it's so lightweight to use the overhead involved in sending out a tweet is 10 seconds or something but the benefit from doing that can be absolutely monstrous so in the lives of busy knowledge workers I think the more frictionless we make participation in these environments the better off we'll be when I look at a lot of the corporate 2.0 technologies I feel so many hoops you have to jump through before you can produce or consume anything but it doesn't, it's not surprising to me that they wither on the vine alright thanks but I don't see companies getting it yet and I talk to them and they're excited and they're open but they're not as we just said they've had to internalize wikis and then blogs and then social networking and now Twitter and those of us who do this stuff full time can very easily lose sight of how weird and frightening this world is to someone who just wants to make dog food all the time so I think they will get there I think the tools will calm down and the management mindsets will get there it doesn't feel like a six month process to me by any stretch Andrew Bestby thank you you've taken a huge amount of information and compressed it into a time that you can spend together so thank you very very much great thanks very much I'm at 99.9