Continue in 2 seconds

Data Warehousing for the Telecom Industry

Published
  • December 01 1998, 1:00am EST

This article looks at the market forces that are motivating local and long distance providers, cellular companies and other communications firms to strongly consider data warehousing as a way to achieve competitive advantage. It also examines how generic business intelligence requirements are uniquely shaped by telecom-specific needs. It also discusses the challenges faced by many service providers as they attempt to implement business intelligence applications and reviews new approaches to data warehousing that have proven successful in yielding tangible business benefits.

The Telecom Marketplace

The breakup of AT&T signaled the beginning of rapid and fundamental changes in the telecom marketplace. Today, the FCC continues to play an active role in deregulating the telecom marketplace by trying to stimulate competition. As a result, communications companies of all stripes are aggressively seeking ways to protect their existing markets while preparing to enter new ones. The ultimate goal is to become a "full service provider," offering business and residential customers a comprehensive, highly customized set of products and services that will ensure customer loyalty while providing new opportunities to generate revenue. For the executives running these companies, this creates a whole new set of business intelligence requirements.

These requirements can be divided into a set of well- defined categories that include growth, financial performance, risk and mix, productivity and efficiency, customer satisfaction, and external factors. In the telecom arena, each area is shaped by the nature of the telecom business model and the realities of the marketplace. This article looks at two of these measures, growth and customer satisfaction, which speak directly to a service provider's ability to gain and retain customers.

Growth

In a generic sense, growth is a measure of a company's ability to increase market share and revenue, including such things as the number of customers, sales revenues, etc. In the telecom industry, this translates into a requirement to track specific growth measures such as increases in accounts, call volumes, new product revenues, etc. However, tracking growth can pose unforeseen challenges.

For example, growth in one facet of the business may actually represent contraction in another. Recently, a cellular provider launched a calling plan that provides rock-bottom, off-peak rates for new subscribers. As hoped, the campaign resulted in a twofold increase in off-peak airtime minutes. This should have translated into a twofold increase in revenue. Unfortunately, the company experienced a net revenue loss, since many of the customers attracted to this offer turned out to be those who were most likely to default on their bills. And the remainder consisted largely of existing subscribers who defected from the more expensive calling plans in order to enjoy the cheaper rates.

Aware of such problems, service providers are investing in data warehouses and business intelligence systems that will reveal the relationship between growth measures and customer information. In the latter example, it might have been useful to analyze the existing subscriber base to develop a demographic profile of profitable, long-term customers that had previously responded to such offers. Then, the marketing campaign could have been targeted only to prospects matching this profile, rather than to the public at large.

Growth can also pose unexpected data management and data integration problems. For example, consider the case of a long distance provider that decides to acquire the cellular firm described earlier. It is highly unlikely that the operational applications used by the cellular provider to track airtime minutes will be compatible with the equivalent systems used by the long distance provider to track long distance calls. Integrating these systems to provide a unified view of growth across product lines may be quite challenging. Consequently, service providers are deploying data warehouses as a way to store data from multiple operational systems in a common format. Potentially, this will provide executives with a consistent way to access information on the performance of different products or business units.

Customer Satisfaction

As deregulation throws markets open to competition, service providers realize they must serve customers better or risk losing them. Cellular providers are especially vulnerable to churn ­ the tendency for certain customers to shift rapidly from one service provider to another in response to small changes in price or other incentives. Since the costs of winning a new customer may be two or three times the cost of retaining the same customer, service providers realize they must do everything possible to motivate customer loyalty.

At the same time, it makes little sense to expend resources on customers who are not profitable just to retain them. For decision-makers, this places a new emphasis on customer knowledge. After all, knowledge of customers and their requirements is among the greatest advantages a service provider has over its competition.

This means being able to answer questions such as: Who are my most profitable customers? What products are they buying and in what combinations? Which are my most profitable products and why? Are my price and service incentives attracting and retaining the most profitable customers?

Potentially, a data warehouse can help answer these questions and many more by enabling executives to analyze and segment customers into groups by their product usage patterns, demographic characteristics, etc. This can arm a service provider with the knowledge needed to fine-tune how its services are packaged and priced so they more closely match the requirements of a particular customer segment.

With customer knowledge comes the requirement for providers to improve business processes that determine how well they are servicing customer needs. Data warehouse applications can be effective if they provide the capability to measure customer care goals against customer care performance. Service providers taking this approach can begin providing proactive customer service, rather than reacting only after a problem has reached critical dimensions or an opportunity has passed.

For example, a provider may want to measure the percentage of calls completed versus calls attempted to see if there is a service problem that could result in dissatisfied or lost customers. If there is a problem, the application should provide the capability to see if this is part of a trend that requires immediate attention or merely an anomaly.

Proactive customer care also means developing concrete ways to measure how the different facets of the service provider's business operations come together to create an experience that promotes customer loyalty. For example, it may be useful to analyze the reasons why people activate and deactivate service in order to identify potential customer service problems and correct them. A long distance provider losing customers because they say they cannot get connected might wish to determine if this correlates with information from the switching network indicating an increase in network faults.

A New Approach

Many service providers are beginning to approach business intelligence requirements from a new perspective. Instead of trying to build data warehouse applications internally using generic tools and technologies, they are buying telecom- specific business intelligence applications focused exclusively on the business intelligence needs of telecom decision-makers.

Such systems combine the functionality of a traditional executive information system with the analytic capabilities of a decision support solution. They give the executives an immediate high-level view of their company's performance, while retaining the capability to drill down through multiple levels of detail to identify problems and pinpoint opportunities.

This is accomplished by distilling the business intelligence requirements into a set of key performance indicators (KPIs). These are specific business performance measures that closely match the way decision-makers think about their businesses and the thought processes they employ during the problem-solving process.

The KPIs must be explicitly focused on the company's products, services, markets and organizational structure. They must relate to business functions in which a particular person or department has clear responsibility. And they must be quantifiable to allow executives to measure, compare and analyze how their companies are performing.

Typically, each KPI will have multiple dimensions that let the decision- makers examine their business in a meaningful way. For example, it may be important to view sales revenue over a specific time period, regionally by service territory, by type of product or customer segment, or using a combination of all of these dimensions.

The executives may also want to view the key performance indicators using different units of measurement. An executive at a cellular operation, for example, may want the capability to analyze network usage both in terms of total airtime minutes and the tariff dollars these minutes represent. These key performance indicators may be updated on a daily basis ­ or more frequently if needed ­ to ensure the information is always timely, accurate and actionable.

The providers of these applications must have considerable experience in the telecommunications arena, which is reflected both in their products and their organizations. This is important because, during the implementation process, the vendor will provide not only technical personnel (such as data modelers and database designers), but also telecom subject-matter experts who will work closely with the heads of marketing, customer care, finance and other business units to define the key performance indicators and identify other business-critical application requirements.

Back To the Future

Growth KPIs may include measures of enhanced services utilization, call volume and duration, etc. Much of this information can be derived from an analysis of the data contained in the call detail records (CDRs). These records are created automatically by a network switch whenever someone lifts the receiver off the hook and include every aspect of a call. These records, generated by the millions each day, are used by service providers for billing and other purposes.

In first generation data warehouse applications, it was not uncommon for a service provider to load all CDRs in their entirety into the warehouse. This "push" approach ­ put it in the warehouse and they will come ­ resulted in huge warehouses that were rich in data but poor in business intelligence.

Second generation data warehouse applications employ a "pull" model in which the warehouse is loaded only with the detail data needed to populate the key performance indicators and to permit specialized queries by an ad hoc query tool. This drastically reduces the storage requirements for the warehouse. It also ensures that executives can view the contents of the warehouse in a format that is immediately meaningful and useful.

In addition to CDRs, the warehouse must include information from other sources, such as the customer care and billing systems, network switches, etc. This will allow the executives to gain an integrated and correlated view of their business processes. For example, they may wish to analyze how increases in network usage correspond to the numbers of new/lost subscribers or how increased network traffic is positively or negatively affecting revenue trends per product, customer segment, service territory, etc This ability to see how one KPI influences another is among the most useful and powerful capabilities of these second generation business intelligence applications.

Customer satisfaction is affected by a variety of business processes. Some involve a direct interaction with the customer (e.g., the customer service representative in the call center who takes a request for service); others are not (e.g., the network switch that determines whether a call is connected).

Some communications providers are approaching customer satisfaction issues proactively by attempting to improve each of the business processes that impact upon service quality. For example, they are trying to improve customer satisfaction by identifying ­ and tracking ­ each step in the problem reporting and resolution process. For example, a South African telecom provider is implementing a "mean time to repair" KPI that will measure the average time required to correct faults and provide capabilities for analyzing fault resolution by multiple dimensions, including fault type, percentage of faults cleared, products associated with these faults, customer segments reporting the faults, etc. The KPI will also include capabilities for analyzing this information using maps, ranking views, time period and forecasting views, etc.

It might be interesting to know, for example, whether problems with certain products are corrected less efficiently than others or that certain service territories are under-performing when it comes to resolving customer complaints. This provider will be able to analyze its problem resolution capabilities on a more granular level using a combination of dimensions. For example, they will be able to determine how their performance in fixing a particular type of fault for a particular type of customer in a particular region is changing over time, whether they are meeting their service benchmarks and how they are likely to perform in the future if that trend continues

By focusing on end-user requirements, rather than data warehousing technology, providers of packaged business intelligence applications are helping to transform the ways in which telecom decision-makers manage their businesses. In the process, they are ensuring that these service providers have the business intelligence and organizational agility needed to compete effectively in this rapidly changing marketplace.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access