Table of contents The active advantage Active data warehousing brings sophisticated insights to day-to-day business decisions. The moment of truth Getting the most from your active data warehouse requires new approaches to your business. Putting theory into practice Extending the use of your enterprise data warehouse to active data warehousing requires a focus on integration and interoperability. The sky is the limit Applebee's, Continental Airlines, Hudson's Bay Company and Travelocity share secrets of their success. View full document > PDF (1818 kb)

Print	Send to colleague
PDF (505 kb)	E-mail us

The sky is the limit

Extending the enterprise data warehouse provides companies with new tools for profitability.

by Neil Tardy

Companies invest in data warehousing technology because they recognize the benefits of analyzing their corporate data and using this information to more effectively conduct their business. Even when the benefits of data warehousing are apparent, companies seldom realize the new horizons of business intelligence (BI) that will open up for them until they actually begin exploring their data. Sometimes a single question yields a new business opportunity; in other instances, a series of targeted questions can transform an entire operation.

The following are stories of four companies that have used enterprise data warehousing (EDW) to provide “right-time access” based upon their specific business objectives. These companies are finding that active data warehousing is helping their businesses profit in ways large and small that they had not imagined.

On the move
Travelocity’s business is helping people go places. To that end, it’s always looking for ways to make visits to its Web site more productive for both the company and the user. With its United States operations, as well as wholly-owned and joint-venture European and Asian subsidiaries, Travelocity has long been a model for what can be achieved through BI. The company doesn’t just track customer data, it uses this information to enhance the customer experience. Travelocity’s targeted e-mail campaigns bring information about low-priced airfares, hotel rooms and car rentals directly to customer inboxes and at the Web site as the customer is navigating.

“The bottom line is we have a lot of information from customers who have booked with us,” says Michael Hawkins, Travelocity’s director of data warehousing and customer relations management. “We know your origin, where you’re booking out of and what you’re shopping for, so we can get a sense for your interests and quickly generate campaigns. One of the things we do in the data warehouse is that we start looking at the different markets and the fares and the costs of suppliers’ products over time, so that when we see particularly good fares, we’re able to market them to you.”

As effective as the e-mail campaigns have been, Hawkins and his colleagues began to wonder if the company could provide even greater customer convenience. Instead of aggregating the information one night and e-mailing it to customers the next day, for example, what if Travelocity displayed the information to the customer in real time, on the Web site? “I guess it was three of us, one of our vice presidents in marketing and two of us tech folks, talking at a PARTNERS Conference,” he recalls. “We realized that we could do this.”

That conversation took place about two years ago. The subsequent Web project is ongoing, although Hawkins says that it has helped that the data being queried online was in Travelocity’s data warehouse from the start. Teradata applications and utilities have much to do with the smooth transition, he notes.“The technology really hasn’t been the difficult part of it, because from a data warehouse perspective, we already had it. Teradata technology, in particular, allowed us to just open the data warehouse up to these tactical queries that were coming in from the Web site.”

Ultimately, the goal is to provide even greater personalization capabilities for visitors logging on to www.travelocity.com. Hawkins wants Web site visitors to be able to retrieve information about specific airlines or hotel chains and receive unique Web pages, or impressions, of special airfares to preferred destinations.

Imagine, for instance, that you’re looking for a good deal on a trip to the Bahamas. “Because we track fares, we can watch the prices over time and as they drop, create an impression back to you that says, ‘Look, we’ve got a special deal in the Bahamas,’ ” Hawkins says. “We started this with the e-mail side on a batch basis; now we’re starting to do that online.”

The idea is to modify the site during a user visit. “It could be anything from your personal preferences that you put in, to your hotel types and your airlines,” Hawkins says. “We’re certainly not there yet but this is possible because of the data that we have. To be able to act on it you have to have an environment that can return a result across multiple subject areas in real time; Teradata gives us that capability.”

One thing leads to another
Companies are getting benefits even with more modest load rates, for example, in the restaurant business. Each year from 1993 to 2005, Applebee’s International opened at least 100 new restaurants; meanwhile, the company’s Teradata Warehouse is growing at a similarly dizzying pace, which is just fine with Seth Jensen, director of sales analysis for Applebee’s. Jensen firmly believes that all information can potentially yield valuable BI. “We’re not necessarily unearthing value on a daily basis, but the discovery process is quite fun,” he says. “Having a data warehouse enables us to ask questions and perform these discoveries. Then, from time to time, we get something tangible to take back to our operators and say, ‘Look, here’s a way to look at the business that we never thought of.’ ”

Indeed, Applebee’s has learned much by sifting through the contents of its data warehouse—and each new finding has seemingly led to another revelation about the business. Most recently, the company discovered a way to determine the average total time hosts, wait staff and bartenders (known in the industry as “front-of-the-house” employees) spend on a single customer order, from offering the initial greeting at the door to bussing the vacated table.

“It’s a new way to measure front-of-the-house productivity,” Jensen explains. “Before, we always measured it in the classic way that everyone else measured it: our front-of-the-house cost is X percent of sales, so we put X number of people on for X number of guests. But it’s not what you’re spending, it’s what you’re doing with that labor.”

This new business measurement was revealed only after Applebee’s had already used its data warehouse to learn how long its cooks require to prepare each menu item—stemming from an implementation the company calls the Kitchen Display System (KDS). This productivity measurement came about when Applebee’s discovered that it could calculate the amount of time its credit-card customers were spending in the restaurants during the lunch and dinner rushes and subtract out the KDS cook times.

Simply, the KDS is fueled by culinary logic. It takes longer to cook a steak than it does to make a bacon cheeseburger. The steak and bacon cheeseburger both require more time than is needed to prepare a salad. Of course, anyone who’s been in a kitchen knows this, but by assembling and then extracting these details from its data warehouse, Applebee’s can apply this logic to deliver the freshest meals possible for its customers, as Jensen explains.

“If it takes 16 minutes to cook a steak, seven minutes to make a bacon cheeseburger and five minutes to prepare a salad, the KDS system won’t send the order for the burger until the steak’s been cooking for nine minutes, and it won’t send the order for the salad until the steak’s been cooking for 11 minutes. By the time that steak hits the expo line ready to be taken out to the guest, the salad’s fresh, the burger is fresh and they’re all off the expo line, not clogging up our ability to execute at lunch,” he says. “It’s similar to the just-in-time inventory approach to manufacturing.”

The company also uses its data warehouse to better estimate the amount of time a rushed diner might expect to spend at lunch. “It’s important to understand how you are managing the guests’ time. If the average time for a lunch credit-card transaction is 39 minutes, knowing about 50 to 60 percent of our transactions are credit-card transactions, then we’ve got a good, representative sample of [the time] the true overall customer is spending when they come to Applebee’s for lunch,” Jensen says. “Where that shows up to the guest is that they might get a 15-minute table time wait quote if they arrive at noon, but if we were managing and understanding that piece of the business closer, we might be able to say it’s going to be a five-minute wait. They’re more inclined to stay and therefore we might be able to do a larger lunch volume in that 11 a.m. to 2 p.m. timeframe.”

Another data component stored in the data warehouse is the Customer Satisfaction Index (CSI). “By storing and combining the robust data provided through feedback from our guests with sales data, we can more readily identify areas for improving the guest experience and growing sales,” he says.

An abundance of the information Applebee’s finds in its data warehouse is distributed to the managers of each of the nearly 500 restaurants it operates; the remainder of the company’s 1,500 worldwide restaurants are franchises that are just now starting to place their data in the data warehouse. In the case of sales and productivity measurements, Applebee’s periodically ranks its restaurants by category. Once managers receive the reports, they can contact managers at other restaurants to share best practices and ideas for improving customer service.

“What’s new is our ability to identify who’s doing well and who’s not doing so well,” says Jensen. “It isn’t meant to be a hammer; it’s more meant to help individual restaurants identify areas to share knowledge and improve performance.”

Keeping things honest
Retail companies lose billions of dollars annually due to theft and fraud. But at retailer Hudson’s Bay Company (Hbc), a detection system is paying for itself by reducing fraudulent returns. “The basic principle is that it’s fraud detection at the point of sale so we can readily detect illegitimate returns,” says Mary-Jane Jarvis-Haig, Hbc’s senior manager of BI.

The Hbc Returns Control Application combines a Teradata Warehouse, BI software and IBM WebSphere middleware to create an electronic link between customer purchases and returns activity.

When an item is returned to any of Hbc’s more than 500 locations throughout Canada, an associate scans the merchandise at a point-of-sale (POS) terminal, which causes a query to be sent to the Teradata Warehouse for a corresponding sale. If no such sale is located, the return is flagged.

Because not every return that gets flagged is an attempted fraud, the Returns Control Application includes override functionality. Jarvis-Haig notes that, because it takes 20 to 40 minutes for each Hbc sale to be posted to the data warehouse, the application could flag, for example, a legitimately purchased item that a customer decided to return before leaving the store. False positives could also result when customers attempt to return items that they purchased several months previously. “One of the delicate parts of the process is when the associate must suggest to the returner that their return is not legitimate. But well-trained associates know how to ask the right questions,” says Jarvis-Haig. “Keeping the overriding under control is a matter of education, monitoring and ensuring compliance to the proper procedures.”

It took only six months upon launching the Returns Control Application in late 2004 for Hbc to recoup its entire investment. The company has confirmed that the number of attempted fraudulent returns has declined since the application was implemented.

Successfully executing active data warehousing requires more than just writing some SQL. An organization needs to define “active” based on its service level agreements (SLAs) and only implement active data warehousing for applications that will truly benefit the company and provide return on investments (ROI). “Every company has to decide what active data warehousing is going to mean to them in terms of investment, in terms of system integration, in terms of data integration,” says Jarvis-Haig. For Hbc, addressing returns fraud was the application that drove them to near real-time data access, while other applications, though not real time, still provide value. “Right now, with the Teradata Warehouse, our merchants and buyers are able to query their sales and their inventories next day,” she says. “It’s pretty immediate.”

Flying high
In some organizations, the data warehouse is like a private playground of information where only the technical designers in IT and perhaps some selected sales executives are allowed to conduct queries. In stark contrast, Continental Airlines is reaping the benefits of its more-the-merrier approach when it comes to data access.

Anne Marie Reynolds, technical director of data warehousing for the carrier, explains that Continental not only encourages but also actively trains its user community to sort, mine and analyze the corporate data. “We may be unique in that we try to push out the analysis to the actual business users more than I believe some other companies do,” she says. “Once people get proficient, they’re free to do whatever they think might help them do their job better.”

The company doesn’t make a distinction between real-time and traditional users, she says. “We acquire our data sources with the shortest possible latency, taking into account the capabilities of the system feeding the data. The users understand the latency for each data source and are sometimes able to take advantage of the fact that it is available real time.”

More than 1,000 users from within Continental have access to the company’s data warehouse. With those capabilities, departments such as human resources, payroll and even cruise-scheduling are uncovering a wealth of useful information. “A lot of times I don’t even know what the data warehouse is being used for because people are doing things themselves and they don’t need us to help them, which is a great thing,” says Reynolds. “It really leverages the investment.”

The active data warehouse has allowed Continental to establish an intricate revenue-management system. Each flight is analyzed and seats are assigned to booking classes. The seat inventory by booking class is updated as seats fill and as the departure date approaches. “Revenue management is a real-time operation,” she says. “We want the plane to be full, but we also want to maximize our revenue per passenger. We are always working to decrease the latency of the various data sources that are used as input to this process. This helps our revenue management users make better decisions about when to open and close the different booking classes to optimize the revenue.”

In addition, the airline is using its Teradata Warehouse to help ensure that its best customers receive first-class upgrades. If first-class seats for an outgoing flight are unsold, the system accesses customer-value information to determine the order in which customers should be upgraded. “One of our big focuses is making sure that our first-class seats are always full,” Reynolds says. “We’re able to use the data warehouse to enable and monitor the upgrading process so we make sure that we upgrade people in the right order.”

The company has also used its data warehouse to better understand missed connections and the subsequent impact on passengers. “There are a large number of different data sources that we have to pull together to do that analysis,” says Reynolds. “Seat assignments are coming from one source, the actual flight times are coming from another, the booking information is from a different source still. The analysis is really bringing together a lot of information that would have been impossible without the data warehouse.”

These are just a few of the tangible business benefits Continental has achieved through its use of data warehousing, but Reynolds acknowledges that there are intangible rewards as well. “It’s always exciting when you talk to a user who figures something out or comes up with a whole new question that we’ve never thought about before,” she says.

The examples of these four companies illustrate just a few of the ways enterprises in vastly different markets are using BI to enhance their performance. By leveraging their EDWs or extending capabilities of those EDWs to perform active data warehousing, companies are acquiring the right information at the right time to improve their strategic and operational decision making.

It’s not enough to have the active data warehouse, of course; you’ve got to have a company prepared to leverage it. Getting support from the business side of the company is the challenging part, Tavelocity’s Hawkins notes. “It really wasn’t just about the technology. That was the enabler that allowed us to do this. I think the real challenge is the culture and the relationships and trying to put all of the pieces together, because you have so many people with so many different world views,” he says. “The technology, once we made the decision to go for it, just kind of fell into place.” T

Neil Tardy is a freelance writer who’s covered business technology for the past nine years. He can be reached at editor_njt64@yahoo.com.

Photos courtesy of Applebee's and Continental.

© Teradata Magazine-June 2006

RELATED LINKS:
Right-Time Business Intelligence: Optimizing the Business Decision Cycle PING Inc. Getting a handle on data Hudson's Bay Company conquers the information wilderness Hudson's Bay Company Best practices in ROI Continental Airlines ECS Executive Summary The chronicles of active data warehousing

back to top