Senior vice president and chief technology officer, Orbitz Worldwide
Roger Liew is, first and foremost, a developer. His working life began on the Web, coding the Web presence of iconic brands such as Victoria’s Secret and cars.com. By 2000, he was working as a developer on the very well-funded startup, Orbitz.com. In fact, he was the first developer Orbitz hired, and he led the team that built the infrastructure behind what is today a publicly traded company responsible for more than $11 billion in travel bookings each year. Now, he manages a team of hundreds of developers, cultivates data to improve and personalize customer experiences, and manages the day-to-day information technology needs of this 1,300-person organization. Here, Liew talks about melding rigorous financial systems with an entrepreneurial, Internet culture.
You cut your teeth as a developer. What does it mean to be the CTO of Orbitz, which is essentially a technology company?
There are a lot of different aspects to it. On one hand, we have traditional CIO responsibilities. We are a public company, so we have rigorous financial systems and reporting needs. But we’re trying to meld this entrepreneurial, Internet culture with grown-up company processes, and try not to let that side of the house strangle the hard-charging side.
The way we’ve approached it is through small teams. I’m a big believer in the power of small teams. So we attempt to make our teams as independent as possible, so they can build things of value on their own. The teams consist of about 10 people, and they are empowered to work on their own backlog of items. They work with the business to evolve the products they own, identifying the features they want to build, and determining the success or failure of those features.
These are agile teams?
We use a process called Agile in the Large. At the base level, the day to day, they operate in a scrum fashion. But beyond that there is a set of practices that dictate how the thing they are working on is defined and deployed, and how it fits into the bigger picture.
You mean the business results of what they’re doing?
We spend a lot of time exposing the teams to the monetary result of what they do. I’ll share a great example. A couple of years ago we were looking at our availability, and we felt it wasn’t good enough. So we determined what we thought the availability should be. But there was a lot of pushback from the developers, and some frustration around prioritizing that work against more innovative things we were working on.
The real breakthrough moment was when we did some simple math: for every hour of downtime, or every minute of downtime, this is what it’s worth. And the moment we talked in terms of hundreds of thousands of dollars, it got everyone’s attention; both the developers and the business leaders.
How did the developers receive this bottom-line logic?
It was almost an instant transformation, the moment we started talking about dollars and cents. People are rewarded in monetary terms. And you only get those rewards if you hit your financial goals. In retrospect it was really obvious, that this is the only way you should talk about availability. But we had talked about it in terms of hitting three 9s or 15 minutes of downtime a month. That’s too abstract. But when you’re talking about millions of dollars being sacrificed to downtime that is controllable, it’s hard to argue against the importance of that work. And we eventually eliminated scheduled downtime altogether. Because for a global e-commerce site, there is no good time for downtime.
For a developer, there is a lot of CMO in you. How did you pick up those marketing sensibilities?
That’s the advantage of growing up with the Web. When I hit the workforce, everyone was learning how the Web was evolving together. And marketing has become this incredibly data intensive, highly computerized discipline. A lot of our investments in data infrastructure are around marketing efficiency. And I don’t think you can get formal training in that area. It’s moving too quickly. It wouldn’t help going to a classroom to try to learn that. It moves too fast. You have to keep trying new things. That’s the fun part of working in a consumer-focused company; there is a lot of experimentation.
All for the consumer’s benefit, right?
Of course. The reason we do the experimentation is that we want to make the experience better. That’s the great thing about the consumer business, as opposed to B2B. Change is hard for any organization. But consumers tend to accommodate change better if they get some benefit.
I mean, we are continuously rolling out new features. All the time. We do close to a thousand code deployments a year. And then once it is deployed, we try different versions of the code. We A/B test every significant feature, and most of the minor features too. That’s how we know if we are making the experience better or worse. And if we’ve tanked it, it doesn’t take long to find out.
And all of this work is presumably leading toward a more personalized customer experience?
Data is the key to this area of opportunity for us. We have been using it a lot on the marketing side, and we have a vision to use it to create personalized experiences for everyone. That’s where we think all of this data will lead.
Remember, in the old days people used to book travel by talking to someone knowledgeable who could tell them where to go and where to stay. And the online travel experience isn’t quite there yet. But we believe that we can get closer to that, so people will have to do less work to plan a trip, and feel empowered to make all the right decisions.
So what’s holding you back?
At the moment, there aren’t enough people that know how to build these things. There is huge demand for people with expertise in data; data scientists. And education is lagging behind the demand, and the technology, which is quickly evolving. These people are in high demand, and are very hard to find. Especially with experience on big data.