San Francisco 2014
Transforming to a Culture of Continuous Improvement
We started our technology transformation about 2 years ago. It started with a move to “agile”. I’d love to share what we learned from the experience and where we are on that journey.
We went from agile transformation to focusing on a leadership development program (which then became an exercise in defining our culture) to a technology strategy story which includes creating a culture of continuous improvement.
We have varying maturity levels within our organization. I’d like to share our customer mobile story as a case study in CD/DevOps and continuous flow. I can also share where other teams are at and how we’ve approached the challenge.
Scaling this is another challenge we are currently discussing. I can share our plans for how we are approaching it and what we have learned so far.
Courtney Kissler, Vice President of E-Commerce and Store Technologies, Nordstrom
CK
Courtney Kissler
Vice President of E-Commerce and Store Technologies, Nordstrom
Transcript
Good morning, everyone. Um, so just to kind of tack on to what Jean said, um, we have had the privilege of getting exposure to the dev ops community, pretty much, you know, Rob's front and center here. He's our, uh, our, uh, kind of thought leader in this space and got us connected to Jean, um, a couple of years ago. And John all spa as well. Who's been a great resource for us. Um, but I think, you know, you mentioned the connection with target. It's like, what I love about this community is you really get the opportunity to share, and it's genuine. I feel like we're all in this to learn from each other, which is kind of rare. So I think, um, just with that, when I get into the presentation, so from a Nordstrom perspective, um, just to give you some context around the company itself, um, we have about 63,000 employees, 1800 of which are, um, in our technology organization, we do about 12.5 billion in revenue.
Um, and so that's kind of the lay of the land for Nordstrom. Um, Jean mentioned, uh, the way that we're structured, probably not shocking the customers at the center of our universe. Um, but prior to really a couple of weeks ago, we were structured the way that you see on kind of the left hand side of the diagram we had on the upper two boxes, that's considered our full price, um, offering we've got our brick and mortar stores, obviously, um, about 118 full line stores just opened our first one in Canada, um, about a month ago. And we have, obviously we have nordstrom.com and then we have, um, an iPhone app and an iPad app in the apple store. Um, the bottom two boxes are really our off-price offerings. So we've got rack stores, um, 162 of those across the nation. Um, as well as, um, Nordstrom rack.com.
And we acquired Haute look in 2011, if you don't know what Haute look is, it's a, um, basically a flash, uh, site. So we've got a website and, um, an app whole look up, but essentially we were, that's how we were organized. You know, everybody's kind of mentioned silos. We definitely, we have those and continue to have those, but a few weeks ago we organized our business leadership. So I have a peer in the business who has accountability for that full price offering full stop. And then I'm his counterpart in the technology organization. So big win. I, we all think from our perspective, our customer doesn't look at us as a channel based offering. They really look at us as a seamless offering. So we're starting to make moves in that direction.
So,
So I'm going to back up a little bit to kind of explain, you know, what was our burning platform and kind of, how did we start transitioning into, um, not only adopting DevOps, but really looking at a culture of continuous improvement? Um, so 2011 was kind of the, the year of the, and maybe we'll call it the year of the sharks instead of the shark visual. But, um, essentially every year in June, our board of directors, um, uh, they have basically an offsite and they pick a couple strategic, uh, topics in 2011. It was all about online growth. Um, they were looking out at some of the, um, traditional, you know, brick and mortar retailers who had kind of said, eh, nobody's ever going to use the web and no one's going digital. You know, stores are fine. And a lot of those companies aren't in business anymore.
And so the Nordstrom board and the executive team said, we don't want to be one of those. So let's figure out how to expand our online offering. Um, up to that point, I mean, we had a website, we, you know, we definitely had a presence in the web, but it was really treated, um, as a steady state investment. We, you know, we'd have capital and we do features, but we do them twice a year. So big batch, couple, couple of releases a year, and that was pretty much it. Um, I'm actually not going to talk a lot about the website journey, if you get an opportunity. Um, we did a case study as part of, um, uh, O'Reilly publication dev ops and practice gives a lot of detail about what we did within the web world. Um, I am actually going to talk about a case study in our customer mobile space, but first I'm going to kind of give you the background around where our organization was at.
Um, so in 2011 we were, um, optimized for cost, which, you know, a lot of these terms are not foreign to all of you in this room. You know, lots of shared services and annual planning cycle, um, waterfall. I mean, that was our delivery mechanism of choice. We did waterfall for everything. Um, big batch releases. You know, I mentioned the website twice a year. That was actually more frequent than I would say the bulk of what we delivered. Usually it was multi-year projects. Um, however, based on that, uh, success criteria at the time, we actually had a great track record, very predictable delivering on time on budget. Um, just tell like a really short story. Um, we, um, had a two and a half year project to rewrite our in-store clientelling application. It's something called personal book, so we kicked it off. And, and I tell you this because the timing actually lined up really well with this story.
Um, so it started in, you know, whatever two and a half years before 2011. So, um, we delivered it in 2011 and by the time we delivered it, it was essentially irrelevant. Um, it was not built for the customer. It was basically built for us. And by the time we put it into production, it, it basically, it, it didn't resonate. And so it was a big wake up call for us as an organization that we had to figure out a way to deliver in this new context. So then we said, what does it look like to optimize for speed? Um, so speed to market speed to value, however you want to define it, but we said, we need to start looking at delivering differently. And so, you know, we did the let's do agile. Um, and we essentially put a bunch of people together, mostly leaders and said, you know, what's out in the industry.
Are there any frameworks we can adopt? You know, we're not the first company to say we should use agile, so let's go out and let's find a way that makes sense. So we did that. We established processes, we established roles. We did all the traditional things that we always do when we're going to make a change. And then we proceeded to, um, implement that change in a very top down method. Um, we essentially said it's a one size fits all. Everyone will adopt the same standard, the same processes, the same roles. And as you can imagine, it, it had varying degrees of success. You know, we had some people who said, of course, I'm going to do it. Cause I'm getting told to do it. And we had some teams say, I don't even know why I'm doing this. Um, and then we also had, um, we have a leader on our team who said, um, we're really good vocabulary engineers.
So essentially we're doing waterfall, but we're calling it agile. So it was, uh, it was a very interesting point in the company's journey. Um, so we did have some thought leaders who said, I think we need to do this differently. And, and what I don't want anyone to walk away with is that we believe in agile, um, the principles, it was just the method in which we went about implementing it. But the one thing we miss that I think was the key missing was the part about it being team lead. So we kind of regrouped and said, what does it look like when we actually engage the teams that are doing the work and ask them to help us understand the best way to deliver value? And so what came of that was, um, value streams. How do we understand the value? How do we understand our cycle time?
How do we migrate to continuous flow and really embrace the improvement kata, um, and figure out how to deliver value, um, and understand how to move faster. So with that said, I'm going to talk about our customer mobile app team. Cause they were front front, basically trailblazing for us. And it really started with the leader that we, um, we put in, basically they had been led by someone previously, we moved this leader into the role and he really understood lean and really understood a lot of the practices and techniques and principles that we were trying to adopt. So prior to him moving into that role, we had our traditional model. So most of our technology teams, our programs are divided into kind of two silos. I would say one is a delivery. So anyone who's basically putting a feature out and then we have another team that's our production support team.
So not even talking about, we also have operations teams. This is within the development teams. We have this divide. So we have a lot of the folks that are actually producing the code are not the ones who actually fix the issues in production. So that whole accountability challenge. So when this leader went in to this team, he basically said, I want to understand, like how do we deliver value? Like, what does it look like in this team? Um, but he didn't start with the principles, like the terms and getting the team educated on their principles. He actually just went in and did it. And so the result was this, which is, it looks real messy, and it's supposed to, um, but essentially he was able to surface, how are we delivering value today? How many teams are involved? Who do we leverage? Who are our suppliers?
How do we move value through our team and through the system basically. And it's kind of hard to read and I'll, I'll show you a different slide and a couple, couple a minutes, but essentially our lead time was anywhere from 22 to 28 weeks, which is an eternity in the mobile space. Um, but this made it visible, which I think is a really key theme around these practices is that before that we knew we were only releasing a couple of times a year, but nobody really knew why. And this really highlighted, um, what was involved in getting releases into the app store and really made it visible, not only for the team, but for our business and for our stakeholders. So once it was visible, um, the team made changes and I say the team because the team owned coming up with what would make sense and really drove the change.
So one thing we did is we broke down that divide of dev versus prod support and we created squads. So we basically put people together to deliver value and put them all together in the same team. We also migrated to continuous planning. So in the previous context, we would have these huge ceremonies, um, called release planning events. And we would pull literally hundreds of people into a room to talk about like, what's the next release gonna look like? Um, so we stopped doing that and instead move to this concept of continuous planning. Um, one thing that kind of carried over into the agile context from the waterfall, um, methodology we were using was this, uh, hardening phase. And we had an in every project, essentially, it was a window between kind of testing, ending and deployment, um, where we were stabilizing the release, but really what it was, was kind of a catch all like if a feature wasn't really done, you kinda, you kept working on it. And essentially we said, no, we should be building in quality upfront. We shouldn't even need this hardening phase. So how do we, um, eliminate it and actually do the right thing earlier in the life cycle? Then we moved to a single backlog of work. So the business worked closely with us to make sure that we weren't just, uh, prioritizing features all the time, but we were actually looking at production fixes and defects and bugs all within the same backlog. So essentially work has work.
So as a result, um, bugs went down and throughput went up. So this is a really good story from their perspective and from our businesses perspective. So that's kind of one window into the data that we were able to, um, to gather and make visible as part of these changes. The other one, this is my favorite one. So our deployment and release frequency. So before, you know, I mentioned, we do releases twice a year. Now we can release monthly and frankly we could release even more frequently if we wanted to it's. But the key is that the business decides it's not technology. And we used to be the impediment to the releasing frequently. So we've actually found that monthly is pretty standard in the apple store. So we've, we've feel good about where we are, but I think it's just important to know that with this method of how we're delivering, we really could, the business can decide when they want features to go in.
Um, so I'm going to shift gears. Um, we, when we focused on customer mobile, it made a lot of sense, especially when we said we want to be relevant in the digital space. Um, and online growth, obviously having a compelling mobile experience is, is a key to that. Um, but what we learned as we went through this with the customer mobile team, is that a lot of these methods for getting work done applied everywhere. It wasn't isolated just to the places that we wanted to have speed to market. It really was relevant in any team that was delivering value. So one of those teams is our restaurant team. So if you haven't eaten at one of our restaurants, I encourage you to try them. They're fantastic. Um, we have 210 of them, um, because some stores have not only a restaurant, but they also have a cafe and an EMR.
So we've got kind of three possible concepts within a single full line. Um, so this team talk to you about where they were at towards the tail end of 2013. We have this, um, uh, business requests that comes our way, it's called a re concept. And essentially the, um, request will be, I have a cafe in this full line store. I want to rebrand it a marketplace, which sounds really simple. It's like, okay, well change some signage and open it the next day, but that's not really what's involved. There there's menu changes. There are footprints within the actual restaurant where sometimes we'll add servers or registers we'll do configuration changes. So it's pretty involved. So this team did 11 of those in 2013, our business team said we're going to do 44 in 2014. Um, and at pretty much the same time, this team was also going through a pretty significant, uh, pain point around service interruptions.
So they had elevated the visibility of the incidents in the restaurants to what we call a high impact. Um, just to give you a brief kind of how we manage incidents at Nordstrom. Um, once something is categorized as a high becomes very visible across the organization and usually means that, you know, we can't take a customer's money, which is, uh, you know, it's a big deal. So this team prior to 2013, those incidents were not classified as high. They actually were medium. So they really weren't hitting the visibility of, of, um, of a high, even though they really were. So within a two month period, they had 30 plus of these incidents. So you can imagine what that felt like, um, for them and for our stakeholders. And it was very visible. So I got a phone call from the business leader of one of the business leaders in the restaurant division at the end of 2013.
And basically, you know, we're going to open more concepts next year. We're having all these service interruptions you needed to triple the size of your team. Like there's no way that your existing team can keep this up. And basically my response was well, that that's one way to solve it. We could definitely throw more people at it. It's a go-to move for us to be Frank. It's like, we must just need more people. Let's just keep adding more people. And then over time we realized that that's, that's not always the best solution. So since we had just done this, a value stream mapping exercise with customer mobile, it was a little, um, it was, it made it very easy for us to say, why we try this with this team? Why don't we take them through this exercise, let's understand what it takes to do one of these re concepts, and then let's identify the waste and make some improvements.
And let's see if we can actually not scale the team. Let's see if we can do this with the people we have. And, um, and so that's what they did and the business participated with us as well. So it might be kind of hard to see, but essentially, you know, uh, this team was able to take their cycle time and reduce it by 40%, which is one improvement. And, um, it happened to be at the very beginning of the value stream, which if you can find something at the beginning, that's, that's a really big win because it typically has a lot of downstream, uh, payoffs as well. But essentially we were asking for a bunch of information at the beginning of this process, some that we didn't need at all, some that we just didn't need then. And, um, and we were just iterating at the very beginning and it was just creating a bunch of waste.
So the team streamlined that, and they took 40% of the cycle time out with that one experiment alone. And then the bottom graph is to show kind of our trend for incidents. So once this team made the problems visible, um, they also went through the true that technique of problem solving to identify root cause and come up with a counter measure that they basically, um, put together a business case. And we ended up funding, new hardware for a restaurant division. And as a result in the last two months, they've had three, uh, hardware incidents and we actually just completed the hardware rollout. So I'm excited to see what next month is so far. I think we haven't had any, um, so it's a great story. Um, the other thing that's exciting about now that this team is practicing this, we obviously have more work to do cause you can see, you know, there's this other bar where we're also having incidents, but they're for a different reason, but the team was able to quickly get to the root cause and figure out what we should do next.
And the other thing about this team is when I first started supporting them, um, everyone wanted off of the team. They were like all planning their exit strategy because it was so hard to be on that team at the time. And I had a hard time attracting people to that team. It's the same team that went through the value stream exercise. They have stayed on the team. People want to be on that team because they see how they're doing their work differently. And it's actually one of the teams that is kind of trailblazing in this space. So they have been a model for us, um, recognized that some of our, um, annual recognition meetings. And they've been able to really demonstrate that if you look at your work differently and you make things that good things happen, so it's a good success story. So there's some things I want to leave you with.
Um, people, I think you've heard this theme from yesterday too. You know, I talked about a lot of process, showed a bunch of data. That's very exciting, but, um, what it starts with is, is people, if we're not engaging the people doing the work and we're not leveraging what they know and asking them to help, um, I just think it all falls down. So we're a big people oriented company, which is probably not surprising if you know, Nordstrom. Um, but I think it's important for us in our organization to treat our people like we treat our customers. Um, the other thing is everything we build should be about the customer. We should always ask our asks, ask ourselves if the customer would value that. You know, I mentioned that personal book example, this would have been really compelling to do. Um, throughout that project, we should have been asking that question.
Um, I have a passionate belief in continuous improvement. I think it's a critical component of how we'll get work done, um, and really creating a learning culture. You know, I've been in the process of learning a lot of these, uh, techniques over the last year and, um, it's hard and that's kind of the next bullet it's SU it's. Um, it requires discipline. Um, as, as someone who's been at Nordstrom for over a decade, we're fixers, we love to be heroes. We're very optimistic. Um, in this world though, it's important that we are disciplined in how we create, um, value and really also making sure that we understand our capacity and that our teams feel supported and not overburdened. And I think this is a method to get to that. Um, and then leaders have to evolve and I'm going to go into detail about what I mean on the next slide, but I know for myself, um, I've had to make changes in how I lead and it has been a very good, um, learning experience, but again, it has definitely hasn't been easy.
Um, so Jean had asked me, you know, if I could wave my magic wand, which I was like, that would be awesome. I'd love to wave a magic wand. Um, you know, I, what I would want for every leader in our organization and number one is honor reality. We have a lot of, um, assumptions that get made. You know, we put a process in place three years ago, people should just follow it. Um, they just need to understand that that's the process. And, and often there's a reason. There's a reason that people are working around something and it's usually not bad intent often. There's, there's something to it. And I think in this, um, especially in the areas where we're trying to, um, maximize speed to value, we should be inspecting processes on a very frequent basis and make sure that we're not actually creating friction for the teams that are delivering value.
Um, the second comment is about becoming a student and, and go and see, um, this, this has been met with some resistance because some people translate this to micromanagement. Um, we're a big culture of empowerment and I, you know, one modification to this that we actually talked about yesterday is that it's not go and tell it's not go to the team and tell them what to do. It's go and see, ask questions, observe what the team's doing, genuinely engage and ask how you can help, um, versus, you know, traditional kind of sit in our offices and wait for the team to come to us and then becoming a teacher. I think the practice of problem solving and the leader's role in that helping the team get to root cause making it important, giving them the time and the space to actually get to the root cause.
And then following up, making sure that we're helping the team as they continue to go through the problem solving exercise and the improvement kata it's like understanding what it means to pick a target condition and figure out the countermeasures that you're going to use to get to that target condition. Um, and I think it's important to practice it, but then the ability to teach it, I think is, is extremely powerful. I mean, we've talked about leveraging consultants and bringing people in that can help us learn this, but what's been the most powerful is finding leaders who understand it and embrace it and can then go teach it. It just, it has a way more sustainable impact and credit it's more credible. Um, and then leading by example. So most people, when they see the improvements that have happened in these teams, they definitely will say they're in, um, they love it.
They, they want to do it, but because it takes discipline, sometimes the actions don't match the words. So I think it's extremely important that leaders are leading by example and making sure that their actions match their words and then asking why and articulating why. Um, that's probably been one of my favorite experiments over the last year is, you know, somebody will, um, make a statement and I'll say, well, do you know why we're doing that? Well, no, I don't know why we should find out why we should always be asking why. And if a leader can't articulate why then they need to go find out why. And so I think often we take things for granted or we make assumptions, but really getting to the root and asking that question and being able to explain why,
So the, what I could use help with. Um, we have been able to measure throughput and I know I showed the slide from customer mobile. Um, we've been able to come up with some ways of doing that, um, at a team level. We're really trying to figure out how to do that at scale. How do we measure throughput and what's the right metric because we want to make sure that we don't, um, we want to make sure we get the behaviors and the actions for the right reasons. And I think measurement can always be that delicate balance between what you end up getting, depending on what measurement you choose. So that's a challenge for us and then case studies, anything, any stories good or bad? I mean, we're always looking for how can we learn from others and how can we share our stories as well? So anything around that, um, we could use help with as well.
So that's it done?
Hmm.
Thank you, Courtney. We have two minutes. I think that's the time for one question. First one. Yeah.
Um, a great presentation. I'd love to hear more about you shared with us kind of what I'll call more though, a little bit more of a unicorn story as you're talking about getting out an app. Can you tell us a little bit about, um, any of your systems of record or more of those kinds of Nordstrom enterprise systems that maybe more traditional light T how that
Question? Yep. You told us an easy one, but tell us a hard one. I feel about system of record.
It was not easy, but yeah. Um, so we're still early in our journey. I mean the restaurant example is one. I mean, they are not, they are not a, um, I would not label them a unicorn. They're on a very legacy platform. Um, we don't own the code. It's a packaged app. Um, and they have applied these concepts, um, and have been able to demonstrate that they can maximize that speed to value in their context. Um, we have started, you know, I've got someone here from our shared service organization. You know, we've done this in some of those areas, which might not totally answer your question, but, um, like our enterprise service bus team, they've been able to adopt these techniques and actually automate how they provide and provision services. Um, our server provisioning team has gone through and said, how do we understand the value we deliver to these other value streams and then automate processes to eliminate waste. Um, but we're still early in our journey. So we haven't yet gotten to some of the more back of the house kind of teams yet.
Thank you. Point of sale restaurant systems. Easy. Thank you so much, Corey. Thanks.