Host Intro (Gene Kim)

Thank you, Jonathan. To set the context for this next presentation, I want to tell you about one of the speakers, Ross Clanton. I met him at the Velocity Conference in 2013, back when he was director of infrastructure services at Target, where he helped lead the amazing DevOps movement there. He has been part of the DevOps Enterprise movement from the very beginning, helping create and popularize practices such as a DevOps Dojo, Project to Product, and so much more.

I was so happy for my friend Ross when he told me two years ago that he was joining American Airlines, and even more so when he told me that he was going to work for Maya Leibman, who until very recently was their executive vice president and CIO. Ross is now their chief architect and managing director for technology transformation.

During the pandemic, I was so excited that Maya and Ross talked about the technology modernization journey at American Airlines, in one of the industries most disrupted by the pandemic. And as society starts to resume some sense of normalcy, thank goodness, there are few industries that have had to scale from little or no demand to unprecedented demand.

I am so excited that Ross will be co-presenting with Steven Leist, vice president of customer technology, who is responsible for all internal and external customer-facing and airport applications, as well as Julie Rath, who was recently promoted to senior vice president of airport operations, responsible for delivering operational reliability and an improved customer experience to all American Airlines customers focused on airports.

As business and technology partners, Julie and Steven were early champions of this transformation. They worked together to give their team space to experiment with change and ultimately help build the momentum needed for this to become a full-scale delivery transformation across American Airlines. I love this presentation so much that they did in Las Vegas, I wanted to re-air this presentation for everyone here today. So here are Ross, Steven, and Julie.

Ross Clanton

All right. Hello, everyone. Thank you, Gene. I'm really, really excited to be here today with Steven and Julie. They really have been the early champions of this transformation at American long before I was even there, and they helped create the momentum for why we're even here speaking today.

Gene already shared some stats. We are the world's largest airline. We're 25% larger than any other airline out there. Maybe one stat he didn't share is that we have 1,400 planes in our fleet, and we're the most modern fleet in the industry as well. We're also the largest network, and as part of our partnership with the OneWorld Alliance, we can fly people to more destinations around the world than anyone.

But most importantly, our guiding purpose, the mission that guides our whole company, is caring for people on life's journey. All the decisions we make as a company, how we operate, is all guided by that purpose.

Let me spend a little bit of time. I'm going to be the quick show at the beginning, because they're going to tell the real story. I want to talk about where we're at now and give you the broad picture of how our transformation has looked.

We've been on a five-plus-year transformation journey at American Airlines. Our mantra has been deliver value faster. That's been the guiding mantra behind the transformation. There are really two aspects that we've been focused on across the enterprise. One is how we move our whole business into a product model, and the other one is how we improve the agility and engineering excellence in terms of how we're delivering across the whole enterprise.

I joined the company in February of 2020, five weeks before the pandemic. So how have these last few years been? Revenue has been great. Actually, we've been doing pretty good. No, actually, it was pretty crazy. 2020 was a year of disruption, and for us it was about survival. We were going through restructuring, shifting our investments. We rapidly were scaling down our operations as a company, and actually our biggest focus was trying to get our customers to fly again. That's what airlines are doing; we were trying to convince our customers to fly again.

In 2021, we started to get more focused on building that global network, the network I talked to you about a few slides ago. One thing to call out there: we do have the industry-leading global network. But one other big thing we had to do in 2021 was rescale our operations. We had to rehire and retrain more people across our airline than the size of most airlines. I'm talking about tens of thousands of people. Just let that sit with you for a minute. That was the type of scale-up that we had to do.

That brings us to this year. This year it's about accelerating our transformation, getting our business back to a state of normalcy. We have two guiding objectives that guide the whole company: returning to profitability and running a reliable operation. I can tell you, as of September, we've had some of our best operational performance that we have ever had in the history of our airline.

Let me talk about two bigger things we were focused on over these last few years to drive this transformation. First, on moving to a product model: we introduced an OKR model a couple years ago at the beginning of the pandemic, but we had to scale that model for how we prioritized our investments. The challenge was we had enterprise-level goals at a CIO level, but those are broad and not giving enough context to product teams in terms of how they set their goals. Product teams needed that context.

So what we did is we introduced this U Model. It represents the strategic operating layers of our company. Essentially, we said, all right, we're going to set our enterprise objectives at the beginning of the year, and that's going to set guidance for portfolio-level objectives that then gives more context. Our portfolio executives are setting more context around what are the measures that they're going to go after. That informs the product team, so on a much more iterative basis they're iterating through their product OKRs across the hundreds of product teams we have across the organization. Then we build up the right side of that model: the initiatives that we're going to do, the product roadmaps, how we're aggregating those into portfolio-level initiatives in some cases, and then ultimately how they line up with our enterprise themes as a company.

The end result of that, where we're at today, is we actually have 450 objectives across all of our product teams mapped to 750 key results, and almost every one of those is anchored to some portfolio-level metric that they're moving.

One last thing I want to cover, and then I'm going to turn it over to the real show. On the technology side of this, we said, all right, we've got to improve our engineering excellence. We've had a big focus on improving our developer experience, and what we did very early in COVID is we started experimenting with Spotify, with their Backstage open source project, which really is a framework that allows you to build a developer experience platform. We were one of the earliest adopters of that. Since Spotify started that project, we've been one of the biggest contributors in that time.

We've built what we call our Developer Runway. This is built on top of Backstage, and this is a platform that allows us to create, build, and deliver applications much faster, with security built in, operational controls built in, batteries included. You get what you need. Teams leveraging this are able to get from literally weeks to minutes in terms of app creation, which is super powerful. As of today, we have about 35,000 site visits to Runway every month, so the usage of it continues to grow and grow and grow.

Closing out at a macro level for our delivery transformation during COVID, in these two years, two top-level results we track from a DevOps perspective are deployment frequency and minutes of customer-facing impact. We've seen a 35% improvement in the last two years in our deployment frequency, while also getting a 55% reduction in customer-facing impact. We're very proud of those metrics. We've got a long ways to go, but I'm going to turn it over to Steven and Julie and let them tell the story.

Steven Leist

All right. Thank you, Ross.

Julie Rath

Yeah. We're going to tell you how it started and the journey we've been on the past three or four years. And, Steven, it didn't start out so great. There was a lot of finger pointing back and forth, right? I was on the business side, and the team members were saying, "Oh, IT is so slow. They start all of this stuff and they never get it finished. It's never done the way that we want it done." And of course, IT was saying, "We're slow because the business doesn't know what they want. We keep giving them all this stuff, and they're not using any of it that we're giving them."

Steven Leist

Yeah. We had to start using new lingo. In that new lingo, we said, run an experiment, break all the rules, when in fact we really changed nothing. We made no progress, and our team got more and more frustrated.

At one point, about four or five years ago, we hit a breaking point with one of our teams that we had asked to go deliver this great new exciting capability to our customers and to do it in record time. We did tell them, break all the rules. After nine months, we shipped no code to production. No rules had really been able to be broken because the environment didn't allow for it. Ultimately the team was frustrated. Julie and I were frustrated. Our bosses were frustrated. We were sitting in meetings every other week and getting, "You've not moved the dial at all. What's going on?" after about nine months.

So we said, okay, let's meet with the team. We pulled the team together, and actually one of the product technical leads, Tim, is in the audience; he was in that meeting. We said, okay, what do you guys need? What can we do to help you? This is not working. Like all good interventions, do we actually sacrifice a consultant? Everybody feel good? All right, that had to be the problem. Let's send that person home, and now we're going to go win. But we said, okay, we'll do that, but what else do you need?

They said, look, we've been going to DevOps Summit. We've read The Phoenix Project. We feel like this is something that we need to do here. We think we can pull this off, and we think that will be the difference maker. So Julie and I said, we've got nothing to lose. Let's go. Let's try it.

Julie Rath

So what did we do? We took some steps that really had immediate impact on our team members, our customers, and the bottom line of American Airlines. We did some low-fidelity experiments that were quickly tested, but we tested differently. We tested directly with our customers. The business side finally said, okay, it's okay to take a developer out of the technology building. Let's take them out with our customers and our frontline team members so they can see how the real operational world works.

We took our developers right out into our airport operations, into our reservation centers. We had so many great learnings because they connected the dots from the developer all the way to the customer. Our team members could share with developers their frustrations as well, and that went through the whole product team as well. This created a momentum that was really challenging for us to contain, because the team just wanted to do more and more and more.

The bottom line is that we added incredible value because we started to have a number of releases every single day, instead of turning our big requirements documents over to our technology partners, waiting six months, and then saying, oh, that's not what we wanted. This also brought an incredible energy to the teams. The teams got excited, and they wanted to do more. It became a great cycle, with our teams now feeling part of the team and that they were making a difference all the way through to the customer.

Steven Leist

Then we had other teams saying, hey, wait a second, I want to work this way. How can we work this way? So then we had to think about how to scale this new way of working. We brought our friends from Target in to help us pivot to a product mindset. They helped us create our first version of our product taxonomy. We adopted the DevOps mindset. We embraced the metrics from Accelerate. That became our outcomes that we were going to measure ourselves by.

We started to incorporate design thinking and value stream mapping. We started to really create a learning culture, where we were bringing education and knowledge to our folks and giving them the opportunity to learn what it meant to deliver in this new way. We invested in training early with our leaders. We brought Barry O'Reilly in to come teach us what it meant to unlearn and how we needed to unlearn to be able to operate in this new environment.

We implemented playbacks or demos. Every two weeks our product teams have these playbacks, and this is one of the meetings that people don't miss. They show up. We have hundreds of people at a time in these playbacks to see what's going on in our product landscape. Julie and I attend those playbacks. Instead of asking the teams to bring us special status or to come have a special meeting, we go where they are, so we don't interrupt their work. We just get into their flow with them and we embrace where they're at. That also started to make a difference. They felt like we really were bought in. That's where we talked about this public support from us.

Then we also braced for turbulence. They say the only thing that likes change is a wet baby. We knew this was going to be tough, and so we really tried to instill fortitude and perseverance: just don't look back, just move forward. There are going to be learnings. There are going to be frustrations. We're going from a world where we spent nine months trying to break all the rules and we couldn't do it because of our environment. To start working this way and have this success early was really good, but we had to lean in because it was not going to be easy all the way.

Julie Rath

The biggest challenge for us was the COVID response. We had this huge impact to the airline that really took our airline down by 40 or 50 percent with all of our team members. And of course, it always happens in the airline industry: big things happen on important weekends. We got the government regulations and the closing down in the US on Super Bowl weekend. So our Kansas City and San Francisco fans had to have the TV on while they were very quickly developing code to figure out how American Airlines was going to be able to downsize in some countries and get customers trip credits and refunds, as well as all of the testing protocols that our customers needed to know.

This was new for all of us. We also had to figure out how we were not going to have massive lines in the airport, because you had customers walking in with 12 pages of medical documentation with the COVID test. Our team members had to figure out what's an antigen versus a PCR test, and can they fly to this country? It was a huge impact to our operation.

What was best about this process is that our teams got together on that Super Bowl weekend, and in eight hours we had an infrastructure to be able to do this on our digital channels. If we looked a few months prior, that would have taken us three, four, six months to develop. But those team members worked hard with both the business and our technology teams to get that done in eight hours, and it was incredibly complex.

This continued for months and months. Countries changed their regulations. They changed the number of days, and our team had to be incredibly agile to be able to change on a moment's notice. Governments would typically tell us, you have to change the parameters by the next day, but our team was able to respond because of how they came together and how the technology was built.

Another part of the COVID response was that our customers were not comfortable with traveling. If you remember back to the beginning of COVID, it was viewed as if you touched something you could get COVID. Our customers didn't want to touch our kiosks, so they would stand at the kiosks and wait for one of our team members to come and help them navigate the kiosk. We had to respond to that really quickly as well. We really wanted touchless all the way through the customer journey. We did a touchless kiosk experience that a customer could remote in from their mobile app, and we really pushed hard on biometrics throughout all of the customer touch points with us, going to our lounges, our boarding process, and in the future our kiosks as well.

It was a big change, but it was the fastest change that we had ever responded to. Again, our team members felt good about it. That was so important to make our customers feel more comfortable and make the process easier for them as well.

Steven Leist

One thing that happened during this process was one of the industry bloggers, one that's not necessarily pro-American at times, in the comments posting on their blog, said, oh my gosh, American must have this army of elves working in the background for all the things they're releasing during COVID to make it better for us as customers. For us, there was no greater praise or validation that this person took notice that we're doing something different and really accelerating our capabilities, which is really cool.

But again, like everybody here, we talked about this as a journey. We've talked about some of the great things that have happened, and just out of respect to the American Airlines folks, there's about 50 of them in the room today. We get to tell the story, but it's their story. They did the work, and it's an amazing story. But we're not done yet. They know this. We have a lot of work to still do. I'm very proud of what they've done, and certainly it's a privilege for us to get up here and tell the story, but they've done some amazing work.

We're focusing right now on getting better at OKRs. We're in year three of really adopting OKRs around product management and getting better at really trusting OKRs. In some cases we define the work and then develop an OKR, which is the wrong order. We're trying to get better at that.

Prioritization is still a challenge. Being willing to say no to the good so you can do the great is something that we're still struggling with, especially coming out of a crisis. Going into a crisis, we have an uncanny ability to focus. Coming out of crisis, we forget, and it's like we're at an all-you-can-eat buffet and we just keep piling it all on. So we're trying to really say no to good so we can do the great.

We still have some dependency issues in our taxonomy that make it hard, and there's still some friction in our delivery that we need to work through. We're working through the dependency problem. And as Ross mentioned earlier, we're doubling down on our investment in creating this engineering mindset, this engineering culture, and making sure our engineers have the right tools and the right environment to be successful.

Julie Rath

Really important for us is the think big, start small, learn fast philosophy that we've all bought into. It is so important to get buy-in all the way to the top of the company. Steven and I make a point that if we're not agreeing on something, we're going to get aligned before we're ever in front of our team members. They feel that all the way to the top, that the company is aligned and supporting what we are delivering because they align with our corporate goals.

We have to make it a safe place to learn. Our team members have to be able to ask any question. We have people new to our delivery mindset. They have to be able to ask any question they can. We have to coach them, and we have to learn. We are getting great ideas coming from the least expected places, a lot from our customers, from our frontline team members, but also other product teams offering to help another product team. Maybe they've developed some type of code that they think can be redeployed in another product, and they're stepping up and saying, hey, we have some time, we can go help develop that. Everybody is coming together.

Our product owners are a byproduct that we didn't expect. We have a lot of independent contributors who aspire to be a manager of one of our business units, and we have fewer management jobs as you move up, so there's this huge demand. What we've really focused on is the development of the product owner, and our team members understanding they are leading, they are managers. When you're a product owner, you're leading a team, and you're developing the skills just like you would as a manager. Those team members feel proud, they feel like they're developing their skill set, and many people are getting promoted out of those jobs into management jobs now, where they're leading really large teams because they've shown what they can do across these really large product teams.

We also have a big celebration. Yes, it's about the wins. But I remember the very first time we did this, when we were doing one of our first products, we had a failure and there was a huge celebration, cheering, and a win, because we failed fast. So we tell our teams, if you don't fall down and skin your knee frequently because we're trying things and moving so fast, we aren't moving fast enough. We celebrate the successes as well as the failures, but we do have a motto of fail fast, and let's go try it again and move forward.

Steven Leist

Real quickly, we want to talk about how senior leaders can help. If you're a senior leader in the group, or if you need to take this deck back to your senior leaders to get some help, as Erica said yesterday, we believe leadership looks a lot like loving people. We agree. We absolutely agree. Ross talked about our why being caring for people on life's journey. That includes our team members as well.

We certainly believe culture eats strategy for lunch. What I have on the screen here, these are the culture attributes that our product teams have adopted. Our organizations have adopted these as the ways that our teams want to show up and how they want each other to show up. As we do recognition in the company, we'll do hashtag passionate or hashtag selfless or hashtag accountable to help people understand this is what accountable looks like for us, or selfless looks like for us. We can reinforce those behaviors through recognition, but we need our leaders to do the same thing.

We need our leaders to lead with care. We need them to actually create that safe environment. We need to create this culture of celebration, as Julie talked about, and experimentation and learning. We also need them to build in some accountability. But we want it to be outcome focused. Don't tell us how to do it. Just tell us what you need done. Tell us the business problem you need us to solve. We need them to care about how we build things as much as what we build. We can't be asked to take shortcuts on the way we deliver capabilities, because it will bite us in the end. We want to make sure that our leaders embody that.

Julie Rath

The most important thing is that leaders have to set the vision, as we call it, the North Star. What Steven and I have learned is the most important thing is that we get out of the way and let our team go. If we're micromanaging, asking questions all the time, if we just go to these playbacks, a lot of times the questions are answered. Our team is more energized when they aren't managed every single day, and we get better results. That's the most important thing for us, and we really work on the senior leadership model that Steven just talked about.

Steven Leist

We've had the privilege over the past eight or nine years to work under two amazing role models in our outgoing CIO, Maya Leibman, and our outgoing CEO, Doug Parker, who you may not recognize because Ross convinced him to put on a Chewbacca mask, as you do in a video for the company, to talk to the company about how important this effort was. What we call delivery transformation was not just an IT thing. He made it important to the entire company. This is how we're going to change the way we operate in American Airlines, which was incredibly important for us to be successful.

All right, I know we're at time. Just real quick: where we need help. I will tell you, all of the inspiration I've ever gotten for any transformation I've been involved with has been from this community and from the people in this room. We haven't been together in person for three years. This is super reenergizing for me. I need these bright minds to continue to inspire and help us innovate. So collaborating, partnering, let's talk while we're here.

I'll do the shameless plug. I know I shouldn't, but if you are inspired by the purpose of caring for people on life's journey, you can talk to me about coming and working for us. Thank you. Thank you.