In the past, businesses applied group behaviors (sometimes called customer profiles or market segments) in an attempt to predict how any particular individual would react to a specific product offering or marketing approach—usually with limited success. The constraints on data collection, real-time behavior interpretation, and business intelligence techniques made it prohibitively expensive to collect and act on specific individual behaviors.
Given the advances in data storage, computer processing, Internet use, and web technologies, it is now possible to track all of the interactions that a specific customer (or a business subscriber) has with the company. This tracking includes more than just managing billing: It's dynamic interactions based on past purchasing history, calls to the service center, inventory returns, responses to marketing and promotional campaigns, and other sources of customer or company contacts. This information allows individually targeted retention and value improvement efforts and in general permits the development of a more robust relationship with high-value, long-term business consumers.
For the telecommunications industry, this model is a natural fit. Most subscribers contract broadband services, with recurring monthly payments for a extended period. Understanding the needs of these long-term customers is key to maintaining and extending a customer base in a highly competitive environment. Moreover, by anticipating the needs of that base (using previous interactions with the company, such as trouble calls) the company can reduce customer defection and stabilize cash flow.
This article presents the value of developing and maintaining learning relationships with a company's most important partner—the consumer—and offers a viable approach to gathering, storing, and using subscriber life history information.
In any business interaction, there is a certain expectation between the two parties. Each expects something of value from the transaction; each understands that a certain level of cooperation is required to be mutually successful. Neither party wishes to be cheated or ignored. However, depending on the depth of the interaction, there is a greater or lesser opportunity to extend a connection into the future. Let's consider exactly what the term customer means.
A customer is an individual (or company) who purchases a product or service for one-time use. The transaction is simple, and little information needs to be exchanged other than the product or service provided and the price paid. Consider the purchase of a hammer at the hardware store or an apple at the market or even a car from a dealer. In all of these transactions, there is little sharing of information, and the relationship is brief; once the customer has what he or she wants, the customer has no further use for the seller. Even a repeated visit does not establish more than a simple commercial relationship that does little to encourage any form of brand awareness or loyalty.
A subscriber, in contrast, represents repeat business and is a long-term consumer of the company's product or services. With or without a contracted subscription, a subscriber can be expected to continue the business relationship over an extended period of time. Such is the case for most telecommunication firms, where they provide services for a defined period (typically monthly) for a regularly recurring charge.
The difference between these two models of individual interaction with the business should be clear. In the first case, the relationship is transitory, existing only as long as the financial transaction. In the second, the basis exists for a continued long-term relationship. However, as in any relationship, the expectation is that there will be a focus and continued interest on the needs of the partner; no one likes to be taken for granted.
Given the ongoing nature of a subscriber's relationship with the company, it makes sense to consider him or her a strategic partner in achieving success. After all, who is more important in driving company value than your subscribers? So, the goal is not merely to increase your subscriber count but to increase the overall value of the subscribers you already have.
Even when companies say they put our customers' needs first, what do they really know about those needs? Do they know, for example, that the person calling to complain for the third time is ready to bolt for the competition? How often have companies made critical mistakes because they based their decisions on flawed surveys, group interviews, or simple guesswork? Who better to inform the company of the needs of the subscriber than the subscribers themselves? And the interesting part is that they provide this information in every interaction they have with the company, including bill payments, insurance claims, trouble calls, upgrades, downloads, returns, and responses to promotions. The tricky part, however, is that all of these contacts may originate via a multitude of channels (see Figure 1), which creates a significant challenge to piece together all of the individual "conversations" a subscriber may have with the company over time.
Figure 1. Event channel map
Consider the following example. An individual is surfing the web and sees a promotional placement for the company's product. He follows the link and decides to switch from his current wireless provider. He signs up online and indicates that he will pick up equipment (a cell phone) at the counter of a local retailer. Sometime later, the individual gets home and activates the service using the over-the-air wireless programming available from the company. The service is activated and billing is initiated.
In this simple example, there are three channels of communication performing multiple business processes. The response to an online ad, the decision to port wireless service, the decision to pick up equipment directly, the programming of the device (with activation of service), and the initiation of billing—all of these activities provide insight into the subscriber's likely future behavior. Perhaps he is more likely to see and respond to online promotions, a behavior that suggests he may want to receive online notifications of new or upgraded services. Over time, the collection of this information allows the company to make predictions about individual subscriber behavior.
As noted earlier, there are many sources of information on subscriber behavior, but collecting and organizing this information can be an expensive challenge depending on how and where (or even if) this information is stored.
Traditionally, subscriber information tends to reside in a few key databases, such as the billing system, the ordering system, and perhaps the service provisioning systems. The information these business support systems capture is seldom if ever correlated, nor is it readily available for study. Consequently, a great deal of time and effort is expended to "reconstruct" the subscriber's behavior by building complex data warehouses and data marts, which then require sophisticated reports to tie together the subscriber's behaviors over time. This information is often static in that it does not reflect the real-time behavior of a subscriber but only the historical record of some subscriber-company interactions. Moreover, many of the channels, such as indirect third-party sales, are not captured.
In contrast, there is great benefit to the use of complex event tracking, which represents a more direct view of a subscriber interaction. As I described in previous articles (see Resources for links), an event-driven system allows for direct capture of deep system information at the same time high-value business interactions are performed. In terms of a subscriber life history, this approach indirectly gathers business process information that the organization can use both to create a behavior model and to understand the utilization of company IT resources (see Figure 2).
Figure 2. Subscriber event map
Identifying the sources of information on subscriber behavior is only the first step: The next is to organize it in an easily maintained and accessible fashion. Most companies use databases to store critical business information, but rather than provide a physical data model, I have opted instead to illustrate a logical data model, as shown in Figure 3. This model captures the critical linking nature of an account owner (the responsible party for payment of bills and modification of account-level services) to the subscriber, who is the direct consumer of the company offerings. The logical model also illustrates many of the common life cycle elements available to a telecommunications firm. For example, the call records, network utilization, and outage information are readily available in the provisioning systems. Billing information and contact history can often be exported from the company billing engine. Third-party vendors who contracted to provide a service can provide other information, such as insurance claims. In general, this model shows five main categories of subscriber behaviors: customer retention, contacts, service utilization, accounting, and insurance/product repair. There are likely quite a few more interactions, but these represent the bulk of typical subscriber behaviors.
Figure 3. Subscriber life history (logical structure)
(View a larger version of Figure 3.)
This logical model can be readily converted into a number of physical
implementations, such as XML documents, database schemas, or software code
objects. The details of each classifier (for example,
PromotionAcceptHistory) is dependent on each
company's business model but will all typically be linked by one or more
unique subscriber identifiers, such as an account number, subscriber
number, or other unique company key.
All of the forgoing discussion has left out one critical detail: cost. It is an expensive undertaking to instrument all of a company's systems to capture and report on individual subscriber behavior. Moreover, there is the hidden cost of ensuring that the data is accurate and collected in a timely manner, incorporating all of the changes that happen over time to an embedded set of technology and systems. Yet, one can extract tremendous competitive value from this wealth of subscriber knowledge.
Effective subscriber-retention efforts are based on an analytical understanding of customer lifetime value, which depends on an accurate picture of subscriber purchase history, positive and negative indicators (such as trouble calls), and potential for future profits. The accuracy of the statistical calculation for the trade-off between enticements and the probability of subscriber defection is determined by the quality of the subscriber's historical data. Retention efforts can often be an exercise in throwing good money after bad by trying to retain or recover poorly "performing" subscribers. A detailed history of all behaviors—beyond just the simply financial—is an important indicator of success for these efforts.
Increasing the value of current subscribers is always more efficient than trying to gather new ones. After all, the company already has a profitable relationship with the individual, who is presumably satisfied with his or her current service. Past purchases, promotion or offer acceptance, and product inquiries are a good indicator of a subscriber's needs, which allows the execution of business rules to determine the kind and timing for up-sell and cross-sell of targeted products. The now classic example of this kind of targeted sales is online retailers such as Amazon.
Many companies have customer satisfaction as a primary business goal, and for good reason. A satisfied customer is much more likely to have return business as well as recommend the company's products or services to others. It is difficult, however, to calculate an intangible like satisfaction.
One way to understand how satisfied a customer is with his or her purchase or service is to conduct surveys. But you can see a better indicator from each business interaction (through one or more communication channels), where the frequency of inquire-to-purchase, the number and kinds of phone calls or other messaging, to the number and kinds of social networking mentions a person might make regarding his or her relationship with the company is important.
Customer value is a statistical measure of the past and future revenue expected from a particular customer; in aggregate, it is a key indicator of the company's overall value. The valuation of a "customer asset" is predicted from their past purchasing behavior and the probability of future purchases. This valuation is often a critical factor in determining how an investor will regard the company as a whole, which in turn directly affects the stock price. Accurate value reporting can only be generated from accurate predictions, which, as this article has presented, is constructed from a complete understanding of individual subscriber behavior.
As noted earlier, it is important to know when to cut a poorly performing subscriber loose. Subscribers with a consistent set of behaviors that indicate a low value score (for example, late payments, frequent service calls, high activation or deactivation levels, or even fraud attempts) are at high risk for revenue loss and should be appropriately tracked and managed or even abandoned and barred from future service.
A little knowledge may be a dangerous thing, but a whole lot of knowledge is power. As outlined in this article, establishing a long-term relationship with the company's subscribers is a powerful use of knowledge in that it allows well-founded understanding of how a company is perceived by its most important partners. Capturing, recording, and organizing a subscriber's life history are a critical element in gaining that knowledge, which can in turn lead to better, more responsive interactions with subscribers, ultimately leading to increased profits and reputation.
- Check out this two-part series on the enterprise event management framework
(Benjamin Lieberman, developerWorks, 2011): Part 1 introduces the event framework for telecommunications,
while Part 2 provides a prototype implementation.
- Explore developerWorks resources for IBM
Cognos business intelligence and performance management
- Visit IBM developerWorks
Industries for all the latest industry-specific technical
resources for developers. Find more resources specific to Telecommunications.
- Join developerWorks on
Twitter to follow developerWorks tweets.
- Listen to developerWorks
podcasts for interesting interviews and discussions for software
- Stay current with developerWorks
technical events and webcasts.
- Get involved in developerWorks
community. Connect with other developerWorks users while exploring
the developer-driven blogs, forums, groups, and wikis.
Get products and technologies
- You might be
interested in our IBM Cognos Business Intelligence Developer Edition V10.1 trial
- Find other evaluation software by industry.
Ben Lieberman serves as principal architect for BioLogic Software Consulting. He brings more than 10 years of software architecture and IT experience in various fields, including telecommunications, airline travel, e-commerce, government, financial services, and life sciences. Ben bases his consulting services on the best practices of software development, with specialization in object-oriented architectures and distributed computing—in particular, Java™-based systems and distributed website development (J2EE), XML/XSLT, Perl, and C++-based client-server systems. Ben has provided architectural services to corporate organizations, including EchoStar, Jones Cyber Solutions, Blueprint Technologies, Trip Network Inc., Galileo International; educational institutions, including Duke University and the University of Colorado; and governmental agencies, including the Mine Safety and Health Administration. He is also an accomplished professional writer with a book and numerous software-related articles to his credit. Lieberman holds a doctorate in biophysics and genetics from the University of Colorado, Health Sciences Center.