Exploring the Cyber Frontier: Travelocity

Moving toward an active data warehouse, Travelocity grows its enterprise and stands head and shoulders above the rest.

Travelocity, the sixth-largest travel agency in the United States and the largest in Europe, has taken travel—and customer control of travel plans—to new heights. In addition to its U.S. enterprise, the company has wholly owned and joint-venture European and Asian subsidiaries, associates and sites in 12 languages, each delivering the highest-value travel opportunities available.

For seven consecutive years, the company has been named the world's leading travel Internet site by travel agents from around the globe as part of the World Travel Awards. Travelocity's efforts have garnered much attention, with both industry observers and customers rating the company highly. It is its customer focus that has guided Travelocity's use of information and Internet technologies to place travel options on the customer's desktop. At least part of the company's success lies in its innovative use of the enterprise data warehouse, which helps them with marketing offers, online "personalization" and exemplary customer service.

Accomplishing these goals has guided the company to view its IT infrastructure as a lever, lifting both customer expectations and company performance—with the understanding that neither can rise without the other.

A Teradata Warehouse has been at the heart of Travelocity's operations since 1999. Michael Hawkins, the company's director of data warehousing and CRM since 2003, has used that technology to help drive continued growth, and Hawkins says he sees the vast opportunities that the data warehouse brings to Travelocity. Yet he also notes the enormous responsibilities the data warehouse generates in terms of driving planned, managed change. As such, Hawkins is quick to point out that effective enterprise data warehousing is always a collaborative effort, where every player matters.

Q: Internet-based businesses face many opportunities. But, at the same time, the last few years have sometimes brought harsh lessons, for the businesses and for their customers. How do you ensure that those lessons are turned into better business practices?
A: By being the change agent. (Economist) John Kenneth Galbraith once said, "Faced with the choice between changing one's mind and proving that there's no need to do so, almost everyone gets busy on the proof." By nature, (Internet commerce and the data warehouse) provide the chance for massive changes to an organization, its culture and its technology.

Q: And, one assumes, plenty of chances to "do it wrong."
A: Yes, but done right, the introduction of those changes is almost imperceptible on a week-to-week and month-to-month basis. Business has to have an idea of where it wants to go, and the technical people have to understand what needs to be put into place to get there. Together you have to formulate a plan, craft a solution and implement it. At the end of the process, you're in a position to deliver greatly increased value.

Q: That said, how has the nature of e-business changed the nature of Travelocity's approach to data warehousing? Or, alternately, how has your approach to data warehousing changed the nature of your e-business?
A: Traditional data warehousing sought to put all the data into one place where you could get to it quickly. Most companies would say they're great at data warehousing. Our company is about customer advocacy, about helping our customers find the deal they're looking for. That's the business, and a traditional view would be to see us in the middle with the data.

But over the last couple of years, our focus has been to integrate our data across functional areas in the midst of acquisitions in order to get a better view of the enterprise as a whole. We've worked aggressively to take booking transactions and monitor them in a consistent way within a consistent semantic framework, using a standard interface. The view layer is already developed; we can leverage all of our reports across the enterprise without the need to recreate all the plumbing.

Real-world insight

There's something about online business, about e-business, that I believe is self-evident but that also only comes as a result of the data warehouse's presence in the online space. Unlike traditional retailers, we actually know where/what people are shopping for—even if they don't buy. This allows us to craft offers based on people's real-world interests. Try to imagine a hardware store being able to do that, to reconfigure its end-caps based on your preferences as you pass by. This brings a large lift in conversion and is only possible via a data warehouse.

— Michael Hawkins, Travelocity

Q: It sounds like your internal customers are able to take advantage of the data warehouse's capabilities in a standardized way, much like consumers expect that standardized, seamless approach when they visit your site.
A: It's important to understand the business function of the data warehouse. That's one of the things we've really concentrated on. Past a certain point you're not really running a technical shop—it comes down to the dollars and cents based on the value the business will derive from your capabilities.

That, in turn, has to be guided by the understanding that data quality starts at the source. Which brings in an educational component: Quality starts at the acquisition point for data. This needs to be a company-focused collaborative effort. What you look at is what data we have that we can act on and make money from, basically what has a value and what doesn't.

It's all about communication and visibility. One of our goals was to open up the data warehouse in terms of visibility throughout the enterprise. In terms of resource utilization and data consumption.

Q: This educational process was evolutionary as well?
A: Yes. It changed the way we define projects. Data warehouses can too easily become dumping grounds where all the data is gathered for reporting. The result is that people have to deal with the complexity of the data in order to make a business decision. This takes time and expertise.

Another result was that in many cases the projects that came down through the data warehouse were feature/function projects, without much thought of how to use the data, without really looking at what the business wanted to measure. The relationship between the business and the developers was turned around.

Q: And now?
A: We worked a lot on governance. We made cultural changes. We had, for instance, developers who were full-life cycle: designing, developing, modeling and maintaining all the way through. You can end up with a lot of islands of data that way.

So we reorganized our team toward the concept of an enterprise view, toward looking at the company from a horizontal perspective. We partnered with business and met with various lines on a regular basis, talked about subject areas, evolved toward an enterprise view of our various subject areas. On the operational side we introduced more measurement and monitoring, more quantifying.

Q: The results?
A: Two years ago we did key loads in the afternoon. Now we establish SLAs (service-level agreements)—(asking) "When do you need the data?" We measure against that beginning point, establishing values for every increment the data is late and so on.

We worked hard on our service offerings, identifying specific areas, raising them to the visibility of the business, getting business support by establishing a value: If this is improved by X percent, we derive (a certain) associated value.

One consequence was an increased business desire to spend money to fix and resolve issues rather than add new functions for the sake of adding new functions.

Q: So you also put the company on a "weight-loss" program.
A: Yes, it was a project we called "Jenny Craig!" What we did was to identify every piece of data in the data warehouse. After evaluating the usage statistics we determined that much of the feature/function-based project data was unused and therefore not returning value. Also, our retention periods on certain subject areas were too long when weighed against the value being delivered. The end result was that we archived a terabyte of data.

But the real result was the delivery of two messages. First, we wanted to get out of the idea that the (data warehouse) environment is free. Too often when a business uses a data resource, there's no appreciation of the cost of the resource. But we're in the business of growing value, which meant that we helped educate people that a) there's a cost here and b) we'll help measure the value you derive from the resources you use.

We helped business earn more revenue and made clear that data warehouse decisions were budget decisions, not technology decisions. As a result I had my business partners coming with me, making investment arguments for me: "We're going to produce this and earn this much money from it."

And that brings up the second result. We created increased visibility for the (data warehouse) asset and especially for the fact that it's managed.

There's a communications issue, or component, that's vital. For instance, when there's a problem with a report, all eyes turn to the data warehouse. It's our job to create that visibility, to make clear that we know when there are issues.

Q: Data quality is one of the issues you've addressed most aggressively?
A: We've done a lot of testing from a data quality perspective. Some systems are more rigorous in some fields than others, and it's important to keep that in mind. It's one thing to get the structure right and another to make sure the data itself is adequate or efficient. So we've focused a large effort on the quality aspects, on processes that monitor quality in key fields. For example, we can now identify, in real-time fashion, data that's out of threshold or that has bad values we haven't seen before. And have published these results to the company so people can make informed decisions based on the data available.

Q: And throughout this educational and collaborative process you were taking steps toward …
A: Active data warehousing.

Q: Can you describe what that means for Travelocity? What specific value does it bring to your enterprise—or to your customers?
A: The key to this was the latency involved in actioning the data. We'd already started loading data in near real time through TPump that provided a view of the business as it was happening in some areas. This new paradigm of active warehousing meant actioning our historical data in the data warehouse within an operational system in a real-time way.

The Web is another channel in the campaign management process that has been facilitated by our active data warehousing efforts.

Q: So as data accessibility has increased, what changes have you made in your operations to get the best results enterprise-wide?
A: We measure as much as possible—the number of users, business functions, how often functions were used, the percentage of capacity used. Our philosophy is "If you can't measure it, you can't monitor it. If it has no value we shouldn't be working on it."

(However,) now that we're in active data warehousing, the numbers have gone crazy. We used to average 150,000 queries a day; now it's more like a million and a half queries a day. Plus there's a difference in the applications hitting the warehouse. In traditional data warehousing, you have a bunch of users generating queries; now we have 1.5 million tactical queries on top of the traditional DSS (decision support) queries. So we're learning as we go.

Q: Is your team well-prepared for that learning process?
A: We've invested heavily in technology training. We have seven Teradata Masters on staff, all dedicated to helping our employees. And we all bring a real passion to our work and to the business. When you get somebody who's good in technology and also has a passion whereby they see the business vision, all of a sudden you have a team of closely knit, dedicated individuals working as a team.

But you must also understand that effective delivery requires a lot of people outside the data warehouse group. And this is what's so refreshing at Travelocity; that passion for what we're doing fills the entire organization.

The area of the handoff between technology and business is where true collaboration comes in and where you're really able to see the value of the resource you're releasing. T