Moving Forward with the On Demand Real-time Enterprise

68 Moving Forward with the On Demand Real-time Enterprise

by period and customer segment. In the case of group insurance, it makes sense

to look at the loss profile by (say) employer group.

When the key dimensions implement consistent representations of customer,

product, geography, calendar and other essential entities, then the analytic

applications and subject area data marts roll up into an enterprise wide data

warehouse. This is the base information required. Access to that data in real-time

enables a significant business advantage.

2.5.3 Telecommunications

The telecommunications industry, like many industries, has a great deal of

competition, which results in high customer churn. In this industry, customer

satisfaction and consequently customer loyalty is essential. A low-latency 360

degree view of the customer can help increase customer satisfaction and

consequently reduce customer churn.

The telecommunications industry often develops customer profiles through

mining their call data record (CDR) information. To develop these profiles both

operational and historic data is used. Integrating these two data sources is

necessary to develop an accurate profile. In addition to integrated information the

development of accurate profiles requires up-to-date information. Data mining is

traditionally performed against data created through batch jobs. These jobs take

time to execute so the mining is not done against current data. In a real-time

enterprise, the availability of right-time data enables mining applications to use

current data and to create profiles identifying the current market trends.

These profiles are often used to predict customer behavior. Integrating these

profiles with customer call history enables call center employees to predict what

services a particular customer might be interested in. Knowing what customers

want, enables the call center to better serve their customers. Knowing what

services the customer may want to purchase, enables the call center to target

specific customers for specific offers and consequently increase overall sales.

In addition to using right-time data to create the profiles, call centers need

up-to-date customer histories to correctly match a customer to a profile. If the

customer history is incomplete they may not be correctly matched to the profile

that describes them, and consequently an incorrect decision is made. Matching

customers to an incorrect profile is especially likely to occur in the case of a new

customer.

Since there are times when customers are dissatisfied with services, it is

sometimes necessary to offer pricing deals or services in order to keep a

customer. Combining current customer information with mined profiles allows the

call center to determine the value of the customer they are currently dealing with.