Data Quality Issues in Service Provisioning & Billing

Introduction Large corporations maintain a complex array of databases and data warehouses to record data when a customer orders a particular product or services and has recurring interaction with the corporation. The data are used for many core functionalities of the company, such as billing a customer, maintaining the customer’s history of interaction with the company (e.g. complaints), billing history and history of purchases and usage. The warehoused data are analyzed and mined for customer segmentation, customer retention and acquisition activities (CRM programs), target marketing, advertising, developing new product and service offerings and many other activities (e.g. revenue assurance) that can save and generate significant revenues for the company. Thus, considerable money is spent on collecting, storing and analyzing the data. However, the data collected are often riddled with problems and glitches. Bad data can lead to bad decisions, loss of revenues, loss of credibility and damage to customer goodwill. More recently, there is a pressure on corporations to provide a customer with electronic access to the customer’s account for self-monitoring. Ideally, a customer would want to add, delete or update their order in real time, online. This requires a company to have an accurate and timely view of every customer’s services, products and configurations. It is the pressure to provide such e-services to e-enable the customers that has forced many companies to examine the quality of their customer data and push for automatic flow of data seamlessly across various databases and warehouses. In this paper, we discuss data quality issues as they arise in the process of recording a customer’s data, starting from the point of sale, through provisioning of the service, to the mailing of a bill to a customer for products and services. We cannot discuss specifics due to proprietary reasons, however we illustrate with a hypothetical example. There are many interconnected stages involved such as recording the customer order, provisioning it, tracking customer care issues, computing recurring and non-recurring charges, and finally billing the customer. Typically, each stage is maintained by a different organization, in a different database, with poor communication between different databases. All these factors contribute to data quality issues. Using an illustrative example, we will describe potential problems at each stage and propose possible solutions. Many factors, such as legacy systems and decisions driven by the need for quick “time to market” remedies, make it difficult to solve data quality issues that arise in the service provisioning and billing process, especially in a large corporate setting. The data quality problems and processes discussed in this paper motivated the development of a database browsing and profiling tool that we discuss in a separate paper [5]. Unfortunately, the paper is too long to include as an appendix to this paper, so we have tried to quote briefly from it or refer to it wherever possible. Note: We do not consider the process prior to the placing of an order (telemarketing, sales) or after the bill has been rendered (bill collection etc.). Proceedings of the Sixth International Conference on Information Quality