Media, communications and entertainment firms typically deal with massive amounts of customer data entering their systems each day. However, a large amount of this information is rarely “mined” in a centralized manner, according to data warehousing experts.
The need to achieve a consolidated view of data, specialists say, is a common challenge to most companies especially telecommunications service providers.
“Service providers have an enormous amount of data about their customers, but most companies are focusing on only 10 per cent of that information,” according to Larry Goldman, co-founder and senior analyst for the OSS Observer, a Sugar Grove, Ill.-based market research firm specializing in the telecom software sector.
Companies have a difficulty distilling a cohesive analysis because the information comes from disparate sources such as e-mail accounts, phone call records, mobile phone traffic or even Voice over Internet Protocol (VoIP) transactions, Goldman said.
Hewlett-Packard, recently launched a suite of software products specifically geared for the CME (communications, media and entertainment) companies experiencing such problems.
The company’s CentralView portfolio is modular, standards-based and pre-integrated by HP to enable customers to deploy the tools as stand alone applications or in a combination. The product is now being tested by several companies in the CME sector.
Bill Zimmer, world-wide solutions manager for HP, said the product seven key concerns of CMEs: credit risk control, customer retention maximization, dealer performance audit, fraud risk management, revenue leakage control, subscription fraud prevention, and video usage insight.
“We feel that our customers are using information strictly for billing purposes and not tapping into its full potential,” Zimmer said.
CentralView’s data warehousing and business intelligence features enables users to collect information from various sources, analyze the information and produce reports that can help departments optimize campaigns or programs, he said.
For instance, Zimmer said, the product can help cell phone companies determine if new subscribers constitute a credit risk by analyzing data such as call patterns and payment fulfillment schedules – information that comes from separate departments.
Traditionally, investigation could take up to two months as the billing department wait for the second billing period to inquire about payment problems.
By analyzing call patterns CentralView can help users determine if a “fraudulent” caller might be “trying to make as much calls as possible before dumping the account without paying,” Zimmer said.
The product can also help carriers or media content providers determine what additional products to sell to existing clients, the HP executive said. “If customer contact data indicates the particular customer often uses a cell phone, the company might be able to offer appropriate services and products such as video or music downloads.”ehensive view of a customer’s data is very difficult and expensive for most companies according to Boris Evelson, principal analyst for Forrester Research Inc.
“Firms deal with multiple terabytes of information but much less than 50 per cent us used in a centralized fashion,” he said.
However, except for its specific focus of the CME market, Evelson said HP’s offering is not entirely unique. Other vendors such as IBM, Oracle, Business Objects, Cognos and SAP, all offer software that combine warehousing and business intelligence (BI) features.
Companies that want to deploy data mining tools, Evelson said, must first determine if they are “appropriate for your organization culture or environment.”
He advices companies to develop a BI strategy and architecture that will address “obvious and non-obvious BI stack components.”
In navigating this strategic roadmap, Evelson said organizations must take “baby step” tasks. Each segment of the program must be accompanies with concrete deliverables no more than a few weeks apart.
He said deployment evaluators must pick high value, low cost, low complexity targets to ensure initial success and momentum.
Businesses are also cautioned that business intelligence for multi-terabyte data sets may require different architectures and technologies.
Plan for ten to a hundred fold “explosive” data growth, Evelson said.