 Target's driving growth through all the great investments we're making to re-imagine our store experience. As we continue to open flex format stores, we're focused on three areas, college campuses, dense suburban areas, and urban neighborhoods. We then curate our assortment and our store design to what's right for the local neighborhood, and our guests love the experience. My name is Jillian. I work here in the Rittenhouse area at a nearby restaurant a couple blocks away. I find myself here a couple times a week. It's kind of turned into my convenience store. The fresh fruit and produce is all here right when you walk in. There's a peril right at your fingertips, and then you only have to walk a couple steps away to find everything else that you need. It makes me feel at home, which is really, really awesome. My name is Camina Carpillette. I live in Long Beach, California, and I'm a working mother of two. I like shopping at this Target because it's very convenient. It's right next to where I work, and it gets me in and out quickly. I come to the store at least three times a week. They have the essentials, so things that you run out of quickly, you can just come in here and pick them up. I also can go online and make a shopping list, order it up and just pick it up. I'd like to spend my time more cooking meals and spending time with my family, so I can just run in here and pick up a few things and be on my way to cooking that meal and spending time at home. You know, I work with students at the college, and I get asked, well, you know, I'm not from here. Where can I go? And I just tell them right up the street, there's a Target. I'm a third year at UC Berkeley, and we're at my neighborhood Target store. What's most Berkeley about the store is probably the size, and its location definitely, according to campus, it's pretty conveniently located. With class schedules being so crazy and random, sometimes I'll only have an hour, so it's really easy to just come in and go out back to class. I love most targets just because there's always that cheery bright red, but I definitely love the artwork on here. I think it's really custom and different. I grew up with my mom going shopping at Target, and now with this location downtown in Berkeley, I feel like Target's definitely growing with me. This is Lakshmi Sharma. I'm Vice President, Cloud and Compute, and platform engineering at Target. I'm going to take you through the IT transformation journey that's happening at Target to deliver the best-guest experience and how it is happening. It's happening through continuous learning process, using technology all around, putting agile DevOps, using open sources, all of it using networking, data, all of it in action at scale, and how we are doing it. I'll take you through some of the glimpses of that technology transformation and the business drivers behind it. Some of the pieces I will cover, what's the investment that's behind? What are the business driver? What's happening in infrastructure and operation space? What's happening in connectivity? Connectivity is about networking, data center networking, cloud networking. How is this whole thing getting connected? The video that you saw, I will keep referring to that video in how technology is driving that evolution that you saw here and how data is being used, how analytics is being used to drive the entire experience I'm going to take you through. And how could you help? So what are the direct feedback from some of the consumers of this technology, which is our users in the stores and the team members at the stores? I've taken some of the feedback and brought it here to share with all of you. So some little bit about us. Target is 1,806 stores, 1,806 in United States. 30-year distribution centers continue to grow. 323,000 team members all across these stores and headquarters. We are headquartered in Minneapolis and we have a global team in India. We are the fourth most visited retail website in United States and more than 26 million unique visitors and unique mediums and unique visitors each month on average. So when I'm giving these numbers, think of it as IP addresses, think of it as the data endpoints, IOTN, and multiply them by some number, X numbers to get to your IOT devices. So that's the scale of the data and that's the scale of the network bandwidth and latency I'm talking about and that's why technology transformation is important. So $7 billion. In February, our CEO, Brian Cornell, he announced an investment, capital investment of this much amount over the next three years in various categories to drive the sales, to increase the market share and to increase the guest experience at all levels, at digital levels, at mobile phones, stores, all around, all across the sectors. So all of that and the thing that is driving it the most is technology and our supply chain. So innovation in those two areas, innovation and digital, that's all is driving and that's where the investment is happening. And what are the pillars that are going to be used in order to kind of fuel this growth and get us to the market share that we are looking at? Digital first and three pillars. We'll be looking at stores, digital channel and supply chain. How do we bring them? How do we bring them together and make them a smart network? A smart network from a supply chain perspective and smart network from connectivity perspective and what else do you put on top of it to drive the efficiencies across all of them to get to our vision and get to our goal of high scale, high efficiency and the best user experience, best guest experience. So we have five priorities that we have identified and not all of them may sound to you directly as some technical priorities, but think of all of them somehow connected through some technology, somehow being delivered through software or some hardware experience. On-demand shopping, it's not just online. On-demand shopping means anytime, anywhere, same shopping experience. You have your cell phone and you're going to an online. If you have an app on your cell phone or you're using a laptop or you are in-store using a laptop, you're in-store using a phone, no matter where you are, giving you the same experience, same quality, you being able to ship an item which is ordered from online to your home within a time that you want. You can go to a store and be standing in a line, somebody coming to you and getting your order delivered when you're thinking that you have spent more than 15 minutes in the line. Think of all those possibilities and all of them means on-demand shopping experience. Merchandise categories. On the item side, yes, we need to continue to evolve on what are the categories and what are different brands that we will continue to kind of add to our portfolio. Localization and personalization. Now I'll refer back to the video that you saw. And combining the next bullet, which is small format. So a lot of investment that we have announced that's going to happen in the next three years is in a small format stores or in remodernization of a store. So what does that mean? So when we are going from a big general-purpose store format to a small format, which if a lot of customization need to happen in the inventory, a lot of space consolidation need to happen. You cannot buy the same size store as you buy in some remote area. The similar amount of money and the similar amount of inventory, you won't be operationally efficient if you use the same kind of capacity in say San Francisco City or in New York. So in order to address the local needs of the market, in order to give the personal experience to your guests in the store, the small format stores are going to be key. But what it means is that there is a huge amount of innovation that needs to happen in supply chain. Now you cannot use those big trucks sending big boxes of items. Now you are thinking about really your personalized experience of five items of each type or one item of each type coming to a store. Now visualize, each item has an RFI tag. And each item needs to be placed somewhere to be brought from hundreds of locations to that store. The small format store, they may not have any inventory in the store. So how a distribution center is going to send them in the time so that that small format store, which is sitting in a university, which is sitting at the corner of a very populated densely populated diverse kind of area. So all of that is technology. You are talking about how you pull the inventory same day within an hour from a fulfillment center. How do you deliver that to the guest? You know, ship to a guest. All of that in this new format world, supply chain, software and technology, they all get merged together. While we are focusing on the priorities, it's really important that we continue to look at transformation in a way that we continue to simplify our cost. We call it N minus 1, N and N plus 1. So every time we are looking at N or N plus 1, we will continue to kind of look at how do we retire or how do we consolidate N minus 1 technology? How do we consolidate it? Business processes which are N minus 1 before we get to N plus 1. Otherwise the number and the variations of the processes and technologies that we use will be out of control. So connecting with our guests. So target one and a half year ago, like in July 2015, target opened this, target opened house. It's a place of connected device concept. So it's a place where we can invite our guests, we can invite our partners and some customers and vendors. They can touch and feel what an IoT experience looks like. How should we put things on the shelf? How do we use technology to address the needs of this flexible fulfillment we are talking about? The speed and delivery to our guests across different channels we are talking about. So there have been like around 13 months, more than 150,000 guests and customers and partners and even some friends from the retail world like our partners from that industry have also come and they get to experience and feel how technology is put into action. A really good place to see how IoT works. And once we get that input from our guests, when we get that experience from our partners in technology and business, we take that experience to put the best case into use and try to operationalize that into our stores and other experiences. So a good concept. So far we have learned a lot from this and will continue to do so. All the transformation, at any time you are talking about word called transformation, people are the key. So culture needs to be able to support that. People need to be able to support that. Technology transformation needs a lot of investment in people and Target has done all of that. Continuous learning does not happen if you don't allow your team members allow your teams to invest into that learning, invest into that training. So we have gone and invested into people practices around agile practices around DevOps, formal training, informal training and putting special programs in place so that we make it and put it into action in real time. So there are investment in people investment in technology. So when I talk about technology again going back to supply chain putting that format, putting a store like open store concept all of that is about training and continuous learning. So me being there only five months as of this month but a huge investment has already happened here and that's the reason to be honest I feel like we are ready just ready to take that investment that our CEO has announced and put it into action to the right use and to really be a leader in the space. So if this investment has already been happening if the culture is ready what's changing? It's the speed. So if you want to open these many stores just change the way you are putting your format and putting your supply chain being almost real time inventory management on your shelf distribution center or shipping it takes a lot of speed and a focus on that speed and delivery. Linear to smart network what does linear means for just very quick kind of definition from retail like going from a merchandise or a shipping comes from some place X place to a shipping place it goes to some distribution center or some fulfillment center if it is digital then from their distribution center it goes to a store and from store it gets fulfilled or it gets sent to a send to a guest if they have ordered it to ship from store otherwise it gets just directly sent to the guest from a fulfillment center. So think about all of that. So now what's changing from linear model to a smart network means that it can directly go from maybe a shipping place directed to a customer to a technology change. Now when you're those of you like if you're thinking network think of like what kind of bandwidth would be required for those stores what kind of backup links what are you thinking about how do you serve a small store versus a big store you know satellite links all those kinds of concept backup link what would be that like what would be the latency for that be like how does this change it's a technology change as well digital infrastructure we are re-platforming we have re-platform our digital site and we are re-platforming it more and more to support the scale to support the new number of customers and give the agility to our developers so that they can continue to give these applications every day. We have scrum teams that are one day we have scrums that go one day we just put things sometimes in five minutes so that's how truly DevOps we are and in order to support that scale you really need that technology and those people culture and mindset and support re-platforming mobile channels so when you visit a target store you see team members with devices in hand so they can come to you and they can offer you the same online experience as you get from your mobile phone or from your laptop so that needs some platform as well because you are thinking about that being able to we being able to from data center deploy applications to all those 323,000 team members I talked about imagine the scale platform running on those many devices maybe into two into three x devices and then that platform should be able to deploy those applications that I talked about which are being deployed every day so that's the technology transformation I'm talking about more speed, more performance, stability capacity so how much capacity would you need on an Android device versus how much capacity would you need an iPad versus data center the server where you are actually running these applications where you are building these applications so a lot of those decisions to be made but at all the time scale is in our mind that's the scale the number of team members, the number of stores and how fast you kind of move around your items across them how fast you deploy a promotion or a coupon to a guest who is coming online that's all, it's a huge scale we are looking at and all of that would not happen if you don't have a feedback mechanism if an item gets, if you see a shelf which is kind of empty you would feel like oh my god is this store going to close so even from like how do you fast kind of put things on the shelf and if you are bringing experience online experience to a customer how that customer only gets the items and things something which we are all used to doing it but at the back there is some personalization there is some identity management there is some analytics playing the role in order to get you the inventory or in order to get you the items on your screen that you want based on your previous behavior so behavior analysis and the data analytics identity management, they all play a role in order to get you the flexible and the experience that you want for yourself so while we are executing fast we cannot lose the focus focus is on operational efficiency, we make sure that operationally we are right you cannot just deploy an application which doesn't work because when you are getting a sales store we call them pause point of sales or the cash registers that you see, imagining an application going bad at the deploy and you kind of halting all those pause just cannot happen there is no scope for making it happen there is just no scope for failure in those cases you kind of somehow putting an application or an upgrade on digital in imagining digital site going down and think of those 26 million plus subscribers that happen so there is no scope for error when it comes to configuration management deploy, it has to be real time and it has to be right core business investment our core business in digital core and supply chain and the user experience that you get from it that's core business investment store shopping experience I talked about there was an application that just really was launched two weeks ago by our CIO and we were able to save the sale so you go online and then you save all the items there you go to a store, you are standing in the line and then somebody can come to you and they will open your inventory from online and they will just look at this and will be able to serve you so like as if you have bought that into the store so you can just choose your items from any place anywhere online and then you go to any store and you will get that delivered and get that experience delivered of online into a store so mixing both of them, converging both of them and it's all of it is about how you're developing application how you're connecting the inventory in stores where is the data getting stored how fast do you replicate that data what is your data replication strategy so that the information that you put into online that gets reflected onto the cash register where this person is going to be fulfilling your experience so think of replication, think of data here and learn and innovate we are continuously learning and we have a culture of learn if it doesn't work because it's like we are talking about a day cycle we are talking about scrubs running for 2 or 3 days so it's very fast you learn, innovate and you continue to drop the things that don't work and then you just move on with the things that are working so what we have done so far is we started with the information so we started by reorganizing teams what it means is that we are very outsourced focus so we just turned it around now we are insourced we have a lot of engineering talent the ratio has reversed we used to be around 70% off-shore outsourced to 30% insourced number has changed so a lot of hiring in the past 2 years a lot of people who understand all this technology is brought 50-60 PhDs they are data scientists and the artificial intelligence category there are a lot of people like author of spinnaker people who have authored like dockers from all Netflix and Google and Facebook or Cisco or Brocade so we have people from all over working on this technology in order to support this growth very laser focused so the company used to run in a services model which is like project based model it went from project based model to say product model so we have used DevOps agile in order to drive the products use data and continue to use the data to focus on priorities how we operate we have product model agile DevOps we work like and continue to reduce our technical debt and all the newer applications they are using like new technologies like containers and CI CD newer CI CD pipelines newer way of networking all of that so how do we deliver to that experience application and data so everything that is our APIs they are built using every application that has to be written using APIs as the interfaces externalization which is like public facing APIs as well as internal APIs data is all there and then we put layer on top so that you can consume that data so data is a service being offered to any application that is written risk profiling of application and data continuously cloud and compute we are into public cloud use private cloud private cloud based on open stack Kubernetes all of that so all the choices we have all those options we make decisions every day in how do you redesign the application and what load goes where and what is the load balance and through the global load balancer like internal products for the internal load balancing all of that in action continuously use connectivity to give a consistent user experience faster response time to over network all the network segmentation that you that you are aware of very importantly application based VPN when you try to rewrite an application and try to put some components of that application in your private cloud for some reason and some part in public cloud you are thinking about think of NFV but in an application layer so now your network becomes key security of your application becomes key and how do you deploy that application how do you make sure of the keys and the security and the identity of the user who is deploying it and the network that is serving it so all of that segmentation and running at application level I call it application VPN so managing keys and managing security at application level which is distributed across clouds and data centers identity becomes identity federation role based access all of that plays a role in there because you think about these hundreds of thousands of customers hundreds of thousands of team members trying to access these applications identity management is very important or any device can be taken outside a store and then how do you kind of manage that identity devices all the devices that I talked about the team members using like iPods and Android so writing applications for all of them the platform that can here it's not just public cloud or private cloud the cloud that is sitting on your android device and a team member that's a cloud for me that's that's a place where I have to deploy my platform but choice of the platform for that device would be different for the one that you are running in public cloud or running in private cloud and we are huge users of open source huge users of open source so the focus is in infrastructure and operations the focus is simplifying the environment using open source standardizing on compute storage and networking when I use what compute in IT word compute storage and the elements around it they are all kind of same we do kind of talk about that hey administrator is a network administrator of computer storage but lot of convergence has already happened at least for us so we are into converged infrastructure journey we have converged like almost a block with compute storage and even the networking layer has kind of come and converged from a software defined perspective into kind of compute so we have already made that kind of journey and done a lot of performance testing and monitoring in what makes sense in the converged infrastructure so I believe in converged infrastructure and we are making that happen so we have automated infrastructure recovery so today like if somebody calls from store before they call from the store there is already lot of reliability, lot of troubleshooting already happened in the background and that has all happened through automated recovery and automated healing I will quickly run through my slides because I am running out of time so we are we have our own api gateway that offers security, key management everything is like continuous delivery have invested a lot if you think about apg for example from google think of an equivalent of that at api gateway level so all of that developed, managed by us, by my teams at target we are a big user of Spinnaker to do CICD deploy pipeline our same pipeline drops images across all the clouds and data centers I talk about Spinnaker pipeline running drone, driving Kubernetes deployment all of that has already happened because the hybrid cloud and the journey doesn't happen without all of that integration so big users of open source like Kafka, Cassandra Spinnaker, OpenStack, Kubernetes, a lot of investment and all of those we have a lot of experts on the team who are working in this and putting that to action so open source adoption to us means like it's a titan control so we can own our product delivery roadmap because we can deploy and we are not dependent for every single key element of our delivery on the vendors and the partners so our teams believe that open source and adoption of open source has given us this CICD and the DevOps like the way we are doing it because we wouldn't have been faster if we wouldn't have built this on our own and have the support of the open source platform that we have so this is a view of without putting public clouds and clouds of how we are connected like stores to data centers to cloud we have three huge data centers all in Minneapolis and we are into public cloud we have our own internal public facing cloud as well that said we use OpenStack and Kubernetes technologies so using data evolution and some of the technologies we are big users of deep learning machine learning, web skill distributed so when you think about stores you are really talking about not terabytes of information you are talking about terabytes of information in some cases almost every day and then how do we store that information what do we drop, what kind of analytics do we put on top of it that's all part of that whole journey I am talking about using data as a service it's a big, big thing for us we are really sure that it is delivering to the promise that we have how you can help so this is a feedback coming directly from the users or from the deployment people at the stores and other places so better utilization into VPN connectivity performance so we run into these challenges when you are trying to connect to different stores or to clouds or data center so if the VPN goes wrong like IPsec VPN or your MPLS VPN there is very less kind of control we have there and now think of these people who are coming from DevOps and troubleshooting experience they just want to know what is going wrong because they want to be able to fix that so there is very less trending and reporting like it just takes hours and hours to get that troubleshooting going clear recommendation so this is just very recent so it seems like we are already talking about 5G but there are still challenges and LTE is the right thing for us or Wi-Fi is the right thing for us at this particular location versus another location still kind of need more help there like from service providers offering managed services to us self-serve offering if you are bringing applications if you want us to use software as a service please make it a self-serve experience for us because we have moved pretty fast in agile and the DevOps world self-serving is really the key how do you consume being very prescriptive in how do we use applications and make it easy to support those services give the same experience today it's a different experience for on-prem troubleshooting versus off-prem or how do we consume on-prem versus off-prem that takes us lot of time two days of kind of training here and one day kind of training there that's lot of time and different kind of focus that requires so a consistent experience for deploying and troubleshooting on-prem off-prem so those are some of the things that I believe like would help our teams so with that like thank you very much