 It's great to be back here a year later, and we're very excited that yesterday Allstate was announced as a gold member for the Cloud Foundry Organization. Thank you. Many of you know Allstate. We're a leader in the insurance space. We're a conservative company in a very conservative industry. It's all about risk. But what you may not know is that Allstate is aggressively moving to embrace the fail-fast culture. Many people think that you have to throw away the old way to bring in the new way. But what I want to share with you today is that a large successful company can continue to maintain leadership dominance while at the same time incubating the startup mentality, culture, and know-how. It's not an either-or. You can do both, and you can do both well. I work for a great company. We help protect people from life's uncertainties now and far into the future. Last year, we handled over 5 million claims. Since we were here last year at the summit, we've moved up five spots on the Fortune 100 and gained another billion dollars in revenue. We're committed to sustaining that leadership, and the culture of innovation is alive and well at Allstate. We're doing great right now, but we know that we need to earn the right to continue to do well and continue to do well in the face of many threats. Since we were here a year ago, it looks like autonomous cars will be here even sooner, and 70% of our business is auto insurance. So like anti-lock brakes and airbags, we'd love to be a big part of that change, but it's still changed nonetheless. And it seems like every day more and more money is going into software startups in the insurance space. And to me, this is kind of analogous to when cars were being manually built and the assembly line came. The companies that didn't adopt to that new process and new technologies failed eventually, and we do not want to be one of those companies. So we must maintain growth and build out our culture of innovation. And so we're moving from waterfall development to agile development on cloud foundry, more specifically extreme agile on cloud foundry, because that's where we see the biggest bang for the buck. We're putting our foot on the gas and accelerating, and so that's what Compose Labs is all about. Some enterprises create separate companies far away from the mothership and then put their new people and new technologies in that new company and then smash it together against the mothership. And that works for a lot of companies, but we decided to go in a different direction. So we created an internal brand called Compose Labs to provide focus to our new way of developing software. And our brilliant platform leader, Matt Curry, will be speaking today at noon, so I really encourage you to attend that talk. He's gonna talk about Compose Labs and our thinking around that. So like any great startup, we started with a small band of believers led by Andy Zittney. A lot of large companies would go hire the alphabet super consulting firms, but we realized at the time that they were trying to transform themselves as well. So it wasn't gonna work for us. So we took ownership and we started to grow. Our mantra was everything changes and this is more about the business changing than just technology. But we had to start in technology to prove to the business that we could run. In January of last year, we partnered with Pivotal Labs across the United States and in Europe. And we started building products that had no integration at all and very quickly started to find early adopter business partners to start spreading what we were doing. And so I lead the technology innovation group and we had been building market tests and prototypes for key business leaders. And so we approached those leaders feeling we had some trust built up and said we wanna do things a new way and you're gonna have to change too. And so they agreed to give us a chance and starting in January of last year, like I said, we moved from no integration at all to integration with some of our most core systems to at the end of the year, we were working on some of all the state's most strategic initiatives. And in November of last year, we were working on 17 products. 16 were built just by all staters and only one we were using Pivotal for help. So don't think that you don't have the talent to do this. You need enough talent to get started. Product mindset, I can't emphasize this enough. It's our biggest vulnerability. If all of this is about building software faster and not building the right things, it doesn't matter. You have to build the right things. So we take the business leaders vision and we quickly move towards a minimum viable product. We grade ourselves with key performance indicators and that's how we dole out more money for the effort. Many companies wanna meet with us and they ask, we started a few years before you but we're still chipping around the edges. How did you guys get so aggressive so quickly? And to me, it really boils down to three things. Number one is to find the right early adopters. We needed to have those people that didn't wanna avoid the risk and wanted to jump off the cliff with us early, both in business and technology. Now Compose Labs is an entire technology effort and a business transformation. Number two, a lot of companies told us that they sent their five best people to Pivotal and they came back and then got crushed. What we did at Allstate was we took my technology innovation group which was about 30 to 50 of our best engineers and sent them to Pivotal to learn the Pivotal way. And so by having that many people and by building that kind of success quickly we were able to establish momentum and overcome resistance. And then third, I mentioned it earlier, we had a very methodical and strategic plan to go very quickly from just chipping around the edges to deep into our core systems. So now we're working on global development. We have teams across the United States and in Europe and in India. And for us to scale, we have to succeed at global development. So what we're looking at is building out large scale applications with multiple independent teams that are integrated with the larger effort. We're working through that now. To me it's not that different from when we were building services for internal applications. You're not gonna have all those teams in one spot but if you try to do it the extreme way where you have co-located teams with a product manager, engineers and designer and have those teams in different parts of the world it's a logistical puzzle. It's a coordination issue but we can do it. And we're looking to have daily stand-ups with the product management of those independent teams. And we're really pushing on this because we cannot be successful if we can't build on a global scale. One of the really cool things was when we got started through last year, every delivery area, so we have six to 10 major delivery areas that all state. You can imagine what our IT budget is. So each one of these areas is a software shop in its own right. But every one of those leaders stepped up and did at least two products last year. And it was really important for us because we needed to make sure that these areas lived it and learned it so that we could plan future work. And so as we moved through 2015, the whole goal was everybody try this, dip your toe in the water because when 2016 comes we wanna go big. So eventually my vision is that will be approximately 70% extreme, 30% scrum and waterfall. But even the waterfall areas are starting to think about test-driven development and tools like Tracker and Jira. Last year we spent about $5 million with our extreme development on Cloud Foundry. This year it's 30 to 50 million and I certainly expect that to continue to grow. This is a photo of one of our product teams. Our developers used to be 20% productive. They were in meetings the rest of the day. Now they're 80 to 90% productive. We have the same look and feel and all composed labs throughout the world. This is one of our training facilities. We partner with Galvanize and do four month training classes. After the first two months, the teams are actually building products under an instructor. And so we're gonna train people until there's nobody left to train or no more products to build. All state is investing heavily in retraining our workforce. We all know how hard it is to get talent nowadays. Since this is the Cloud Foundry Summit, I should say something about Cloud Foundry instead of just talking about methodologies. We have two data centers. We designed and deployed an infrastructure stack to support a multi-availability zone and multi-region deployment of Cloud Foundry. Each infrastructure availability zone is fully independent of each other and represents an isolated domain across all stacks. 30% of our virtualization infrastructure in our data centers is now Cloud Foundry. It's kind of funny when I was mentioning Matt earlier. When we built out our first data center, we thought, great, let's push our Cloud Foundry instance to the second data center, and the team refused. They didn't wanna go through what they just went through unless they could have it fully automated and do it with one push. And so, okay, we waited and they figured it out and now we can push from one data center a complete Cloud Foundry instance to the other in three hours with zero human intervention. Just recently, we had a need to power down some of our physical infrastructure, which, at all state, that would mean a couple months of coordination and planning, dozens and dozens of engineers on a phone call to manage the shutdown, startup, and more importantly, the validation of all the different applications. Well, when we performed this, we were able to have four engineers do the entire power maintenance in six hours, and more importantly, we had zero impact on our customers. So, we've certainly built a lot of utilities and tools to do what I've been describing here. We're working with Legal right now to see if we can't give that back to the community. So, all this great stuff, you'd think it's all rainbows and butterflies at all state, right? Not quite. The resistance is strong. We're an 85-year-old company, institutions like HR, Security, Legal, Procurement, Real Estate, they have a lot of rules. I picture all state in 1931, when the company was created, just like a startup, with all the rule book is like one sheet of paper, and 85 years later, you could fill a law library with all of our rules. But I will say that when people internalize the threats, and they talk to their peers at other companies that are further ahead of us, and they learn more, they start to see the light, and there's no way we would be as far as we are if it wasn't for those great folks and those institutions that helped us. We still have a way to go. After our CTO left, one week later, the memo came down to get rid of scooters in our hallways. But you'll see, well, and then we had a little rebellion. We have signs that say no scooters allow, don't touch the scooters by order of the management, you know, that kind of stuff. But you'll see that there's a portable table tennis set there, and a soccer ball in the background, so we're still in the fight. Speaking of culture, it's obvious that culture is the hardest thing to change in a big company, right? It's not about technology, it's about the culture. And so the way I look at it is that the leadership gets it, you know, we see the threats, we communicate it throughout the organization, we get it. And the new people, and people that are tied to technology like our engineers, they kind of get it too. They're excited about this new way for us of developing software. But our biggest issue is in the frozen middle. So for decades, big companies hired, compensated, and rewarded people to take orders. And so if you're a director or a senior manager in a large company, you really fought to get to that level. And in a lot of big companies, you try to build a little cocoon around yourself so that you can be safe. So now we come in and we basically say, no, all that's, you know, blow the whole thing up. Your job is not to give and take orders, your job is to remove friction for those small six to eight person teams, and they're making the decisions. I know it's not easy for people. So when I'm asked about the frozen middle, I say that at all state, the frozen middle is thawing right now, because again, we wouldn't have been able to make the progress we've made if it wasn't for a lot of those folks coming on board. But don't think you're gonna get everybody. Most of the companies we talk to say that about 30% of their people either don't want to or aren't able to make the journey. One more thing about culture. Culture to me is like gravity. As you try to push against it, it'll pull you back down. And what you have to do is put in an enormous amount of energy, take some ground, hold that ground, look for other opportunities to take more ground, but eventually you pop through. And for us it would be more and more brights of light. And then once that happens, you know, the culture is really changing. So a little bit of a retrospective, I've mentioned a few of these already. There isn't a checklist. There aren't enough companies that are down this road where you can go buy a book and do step one through 10 and then you're golden. You have to play the cards you've been dealt. There's no secret to this. Metrics are sparse. Obviously, you know, we're run by Actuarials so they wanted to know what's the return on investment, you know, what's this compared to waterfall development. And when we went to companies and asked, they just laughed at us. They said, yeah, you know, we were looking at that too, but there was such an upside to what we were doing, we decided we weren't gonna measure that stuff anymore. And so, you know, don't look for metrics. Talk to other companies. You know, when we were trying to push this through the institutions that I mentioned, I don't speak HR speak or legal speak. It's imperative that those peers talk to their peers at other companies to find out what they're thinking is to start breaking through some of that learning. And so, celebrate successes and failures. What I'm talking about more here is that now that we have some of our extreme product development hitting deep, deep legacy systems, it's very encouraging when I see those legacy areas after the initial friction start to come around. They start to feel like they can question the status quo. They can question all the rules. They can start thinking about automation. And when I see that, that's just incredibly rewarding for us and for them. So in closing, I hope if you got anything out of the last 20 minutes, it's that big companies can maintain success and also transform at the same time. It's not easy and the resistance is strong, but it can be done. One by one, all state is living proof that the usual excuses don't hold water. It's only for small projects. It's only for new projects. It's only for innovation projects. We don't have the talent. Our business people would never go for it. One by one, we're dispelling these, and if we can do it, so can you. So thank you very much for your time. Full speed ahead. Thanks.