Real-time connections

Cingular makes Teradata an indispensable part of its business.

by Michael Donovan

If you remember a time without mobile phones, you probably remember a world without data warehouses. Now cell phones are indispensable, performing any number of critical functions.

Much the same could be said of the Teradata Warehouse—certainly Cingular Wireless, one of the world's largest providers of wireless communications, has found it to be so.

The Atlanta-based company went live with its Teradata production system in December 2000 and since then has steadily broadened the range of applications and functions that the data warehouse performs. Today those roles include customer relationship management, financial management, operations management and revenue management/assurance.

That range has further broadened since the October 2003 arrival of Joe Pignatello. With more than 18 years of experience as a telecommunications executive and industry consultant, Pignatello mixes his practical understanding of the data-business relationship with an informed vision of the effectiveness of that partnership in his capacity as Cingular's senior director of enterprise data services. He is driving Cingular toward its ultimate goal: running an active, real-time enterprise data warehouse (EDW) with current, on-demand information and user availability throughout the company.

Since coming on board, Pignatello has overseen the implementation of new data warehouse functions and helped educate personnel throughout Cingular about the informational, operational and competitive benefits the data warehouse offers.

In a recent conversation with Teradata Magazine, Pignatello discussed the various strategies Cingular is employing to achieve its active real-time EDW.

Q: You joined Cingular in October 2003. How has the first year or so gone in terms of moving toward implementation of an enterprise data warehouse?
A: We're moving along pretty rapidly. Basically, when I started here the warehouse was focused mainly on billing data. We've taken on different projects throughout the year and have increased the number of subject areas while also gaining momentum and moving toward an active, real-time EDW. The EDW that currently supports financial, educational and marketing decision support is playing a large part in designing marketing campaigns. It also supports areas of customer service, the IT help desk and a wide range of users with ad hoc queries, with more on the horizon.

I'm not sure you ever complete a project such as an EDW; it's more of an evolutionary approach. There are always new challenges, new opportunities and new developments. In our case, for example, there's the recent acquisition of AT&T Wireless, which brings with it a huge variety of new opportunities and challenges as the companies begin the merging and rationalization processes. But we've made a very good start—introducing the warehouse into new areas, taking on different projects throughout the year and exposing new areas of the company, such as human resources (HR), to the data warehouse and its capabilities. We can do a lot more today than we could a year ago.

Q: Have people in those new areas greeted the data warehouse introduction with enthusiasm?
A: More than that—a lot more. Not only does the enthusiasm grow in the new areas, it grows in other areas throughout the company as well. As we educate the community about the benefits and capabilities of the data warehouse, more and more people throughout the organization see ways that their particular areas can benefit from the information and tools the data warehouse provides.

Q: You mentioned education. Has that been an important part of the process?
A: Of course. Some evangelizing is definitely part of the mix. But I think it's important to note how quickly people throughout the organization have bought into the data warehouse concept. The more we deliver, the more people buy in and the more quickly they buy in. They understand that they'll be seeing things such as revenue protection enabled from warehouse data, with processes from switch-to-bill reconciliation and on to fraud analysis, for example. The data warehouse empowers customer relationship management; in other words, it not only protects revenue from fraudulent use of the (Cingular) network, but also allows us to better manage our customer base, all from data flowing into the warehouse.

Once people understand the concept of the 'single version of the fact,' so to speak, they start looking over the whole of the company, at both the decision support and operational possibilities of the data warehouse as well as the reduction of the costs of maintaining redundant data repositories.

As an example, when people begin understanding that the data in the warehouse is enriched by its ability to integrate different subject areas throughout the business, there's an increasing demand from people who want to data mine, aggregate and correlate the data within those areas.

Q: So the data warehouse process reinforces itself as it extends throughout the enterprise?
A: Absolutely. Once it proves itself in one area, word gets around, and people in other areas of the company want the chance to see what it can do for them.

Q: Teradata is well-known for how closely it works with its customers. Have you found that to be the case?
A: Undoubtedly. They have been very good partners. We work very closely with the onsite account team as well as Teradata Professional Services. One thing that impresses me is that, as we work together, we learn and grow together as a team. They are there to work with us as we take on new and different approaches to using the Teradata nvironment—especially on the 'load once, use many times' scenario.

The ability to process large amounts of data in an efficient and effective manner, the way Teradata does, allows us to maintain the data not only for the 'single version of the fact' but also for ad hoc queries, analytics and OLAP, for example. Where ordinarily data would be replicated out to a data mart for analysis, we can now create analytical structures within the Teradata environment.

Q: Which provides greater efficiency?
A: And timeliness and savings. The expense of data will increase normally as more data is created and stored within your environment; however, it increases even more so if your strategy includes unchecked propagation or replication of the data. A load once, multi-use environment such as Teradata lets you build the schemas necessary for the OLAP and analytics while still fulfilling the demands of the users and meeting performance requirements.

Q: As a former consultant, you have experience setting goals. What are your goals for the Cingular data warehouse? What's your vision?
A: I think that one of the challenges and a goal is to move the data warehouse out of the Ôdark area' of the traditional decision support center—to expose it to the enterprise and allow the enterprise to see its data in a different light, integrated and enriched. The enriched information needs to be accessible to all people throughout the business, not just the data specialists.

We're in the process of making the data available to anybody who wants it and whenever he needs it. To do that, we're exposing data through a variety of methods: from direct query access on base views to report generation and analytical tools to create a real self-help capability to JDBC connections for operational access to the data. We're also moving toward a more active data warehouse where data will be maintained on a real- or near real-time basis, as well as creating and maintaining master reference data in support of our operational systems. As this paradigm gains support, the processes for capturing the data are accelerating.

The value of data is greatest when it's most current. The quicker you can get data into the hands of people that make decisions, the more effective decisions across the company will be.

But it's important to understand that decision makers are everywhere in the company, from the C-level executives to our marketing and sales people to the customer service reps who may be in the process of helping a customer. With the right data at the right time, better decisions can be made and can create an environment which will enhance the customer service experience for each of our customers.

Q: To the benefit of both?
A: To the definite benefit of all. There's plenty of discovery that happens when data comes together. This results in more efficient operational benefits and more accurate and timely decisions, and it leads to better products, services and support for our customers.

Q: Benefits that are different from traditional siloed storage?
A: There's no comparison. With disparate data repositories, you can rarely be sure that the data between repositories is current or accurate, and those are just some of the issues that the warehouse resolves. The data warehouse, in relation to its intake layer, captures, profiles, provides error checking and maintains the single version of the fact in a main repository at a very granular level. From here, the data can be accessed, queried, aggregated or correlated by anyone through a variety of methods, including operational systems interactions, whenever it is necessary. This is where you start realizing the benefits as well as the great value from a single repository, since you are able to provide a more consistent view of data that has gone through a rigorous process to ensure its quality and integrity.

Q: To more people and areas of the company than would have been possible in those 'dark areas' of siloed storage?
A: Many more. Once the data is available in the EDW and can be accessed easily, there will be more and more people who will want to use it since they can be sure that the data is current and can be trusted. The capability for people who want reports or to do analysis will, of course, be there. However, as people become more sophisticated there will be more data mining, more and deeper analysis, and more use of statistical and OLAP tools. The data warehouse should also be considered as a part of the operational 'grid' as operational systems connect through a variety of connectivity options and access the data from a master reference perspective.

To give an example, there may be people or systems that need access to a particular type of data, like data in a credit table. Today, that data may be in multiple repositories or systems, updated in multiple ways. If you can create the reference data in a single area, like the warehouse, then it can have an owner identified, be updated through a defined process, and then let the data warehouse become the broker for the data through a variety of access or delivery methods.

Q: Again, an increase in efficiency?
A: Yes, but there's something else at work. Too many times, when data is dispersed throughout an enterprise and is maintained in disparate locations and/or applications, the data is usually beneficial only within the particular location or application.

But when data comes together, like it does in a data warehouse, the efficiencies that can be gained are obvious. However, equally important are the processes used in data warehousing to ensure the timeliness, quality and integrity of the data. It can then provide the means to integrate and enrich the data for better reporting and analysis. Many more ideas are then sparked and insight generated when data can be viewed in total rather than separately.

Q: Which, in turn, opens up new areas of business?
A: And, just as important, new ways of using data to drive the business. The data warehouse gives us a good view of our customers currently as well as their history with us. In this way, the data warehouse can be used like a tool that can examine many customer dimensions like longevity and long-term value and, coupled with the current views, determine the appropriate method of customer treatment.

Q: In what ways?
A: In ways that allow us to better segment our customer base so we can take the appropriate action, which will allow us to provide the customer with the best experience possible when they interact with Cingular. This is only one of the ways we can use our data to increase our reputation as a best-in-class service provider.

Q: And the data warehouse in operation enables that action?
A: It can enable that and provide any number of actions off of it. For example, Cingular provides many channels or touch points in which the customer can interact with us like our Web site, our IVR (interactive voice response) or call centers. The customer experience can be tailored to the specific channel, but the data allowing the channel to recognize the customer should be accurate and consistent, meaning that the data utilized should present the same view of the customer regardless of channel.

Based on what we know and learn about our customers, based on how they interact with Cingular or use our products, the data, in turn, can drive a better experience with Cingular as well as support the development and marketing of new phones, services and product areas for the benefit of our customers.

Q: So during direct communications or inquiries, the flexibility of the data warehouse helps attract new customers, enhances the business you do with existing customers and plays a part in how the company interacts with them?
A: No question. A well-run data warehouse becomes an effective business and marketing tool. We use Teradata's CRM product, for instance, to help us analyze our customer base as we design and create better and more targeted marketing campaigns.

We're able to better tailor our campaigns based on the data within our data warehouse, where we get to know about our customer base and what we continue learning about it. How are our customers using their phones? For voice? For data? Do they use certain features more than other features or different configurations? Knowing what products and services they have, and what they like about them, lets us be more accurately predictive about what customers will possibly want in the future. So the data warehouse not only can help support the business, it starts to drive business by allowing better visibility and insight into our customer base.

Q: Even as that data warehouse is driving changes within the enterprise?
A: Precisely. Early on, we started to bring into the data warehouse employee HR data, and we recently brought in data associated with internal education programs. This can give us the ability to integrate the HR data with education data and determine the best education and training for people performing a job function, which could increase their performance on the job.

Q: All because of the increased richness of the data and the data environment?
A: Yes, but it's not just richness. It's the totality, as I've said. It's the data's quality and currency, the frequency of updates, the analytics that are available, data mining, CRM—the whole array of products and applications that work with the data warehouse—and the different methods our people have of accessing the data that lets them perform more effectively.

Q: Which must require quite a bit of management at your end?
A: But the tools are there. We've placed sophisticated measures and metrics on both the frontend and the backend and have implemented a number of Teradata tools, to help maintain the health of our EDW. Teradata Query Manager uses a set of rules that prevent queries which could affect performance from being executed. Database Query Log maintains a record of every query run. Combining this with a metrics-based dashboard, we're able to look at what types of queries take longer to run and what tables are most frequently used and discover the best way to tune the database. We are currently in the process of implementing Priority Scheduler to allow us to determine what communities of queries require higher priorities and adjust our resources to better support them.

Those are just some of the tools that we use with Teradata. We have a number of reporting and query tools, statistical tools, tools that help with data analysis and mining. Business Objects, MicroStrategy, SAS, Hyperion—all of which offer ways to make the data more accessible.

Q: What's on the horizon? What do you see coming for the data warehouse?
A: Before we look at the future, we need to take a look back at the past. Go back 10 years, and you'll see that the primary function of Decision Support Services was reporting. And that was important. To get a report out is one thing, but now real-time data and operational priorities are another. We're finding ways to bring the data warehouse into the operational 'grid,' finding ways to capture real-time, and finding ways to make the data accessible to whomever or whatever needs it, whenever they need it. The drive toward a real-time enterprise, if planned properly and implemented precisely, will become a terrific competitive advantage.

Q: And an ongoing one?
A: The data warehouse has a huge appetite. It's basically capacity-driven, but at the same time, the data has much more value when it's current and can be available immediately. So capacity and performance have to balance.

To a point, you meet the demands with higher-density disks, faster drives and more applications. But the most effective approach is to create balance in the data warehouse. Balance the need for capacity with the need for performance and maintain proper vigilance so that you can predict the demands of the data warehouse.

The capacity curve will ultimately have a tendency to flatten so you need to ask what performance levels you might need to maximize the effectiveness of all that capacity.

Q: That's the single most important goal? Or is it wrong to talk about single goals where a data warehouse is concerned?
A: You have to understand that a data warehouse is not just one thing. The key to it is in the flexibility it has to support many users with many demands. One of our goals is to make our company's data visible to anyone who needs it and have it available 100% of the time. T

Behind the solution

Teradata Warehouse powered by: Teradata Database V2R5, 96 nodes across 5350/5380 NCR Servers Users 3,000+ Storage: Total Disk: 83,808GB User Disk: 38,342GB Operating System: UNIX MP-RAS Teradata Utilities: Teradata Load and Unload Utilities, Teradata Utility Pak Tools/Apps.: Teradata Manager, Teradata Dynamic Query Manager, Teradata Priority Scheduler, Teradata SQL Assistant, Teradata Analyst Pak, Teradata Warehouse Miner, Teradata CRM and products from Business Objects, MicroStrategy, SAS, TIBCO, Informatica and Hyperion

Journalist Michael Donovan has served as a consultant to various companies engaged with information technology and communications.