SPSS - Customers

Hamburgische Electricitäts-Werke AG (HEW)

Situation

In the deregulated energy sector, former monopolists are using business intelligence to occupy niche markets.

The deregulation of the German electricity market in 1998 caused a shake-up in the energy sector. Stripped of their position as a bureaucratic monopoly, public electricity suppliers suddenly faced competition characterized by comparatively low prices. Companies seeking to survive this price war unscathed needed to explore new ground. A case in point is the German power company Hamburgische Electricitäts-Werke AG (HEW).

Challenge

This Hamburg-based company is playing its strongest hand in its competition against cheap suppliers by providing first-class service and advanced customer care. HEW has focused on its for many years. Today, its strategy is to find niches in the private customer market where top-rate service counts for more than any price disadvantages. The energy supplier has begun by offering all Hamburg households value-for-money electricity and a comprehensive service package via HEWfuture. HEWfuture customers automatically benefit from price reductions, shorter contract times, and simplified price structures. Interaction between supplier and customer has been enhanced via a call center and the introduction of a customer card for private companies.

These interactions, in conjunction with questionnaires and power consumption rate measurements, have generated a flood of customer data that the company collects, analyzes, and then strategically applies to its new marketing approach. This approach uses customer segmentation as the basis for direct, target-group-specific customer care.

Solution

First, HEW built a data warehouse and selected a suitable analysis tool for the segmentation. The firm found what they were looking for from SPSS Inc.—PASW Modeler precisely met their requirements. PASW Modeler displays the entire data mining process on screen, allowing users to intervene at any stage during the process.

Interested in Hamburgische Electricitäts-Werke AG (HEW)? Download the PDF

Our search for a suitable model has shown that PASW Modeler analyses... provide an important aid in deciding which data provide us with the information we require for CRM purposes.
Peter Goos
Sales and Marketing
Hamburgische Electricitäts-Werke AG (HEW)

This was not the Hamburg company’s first introduction to SPSS Inc. The electricity supplier has been working with software developed by SPSS Inc. since the 1970s. What swung the decision decisively in favor of PASW Modeler was the fact that, unlike other evaluation tools, software from SPSS Inc. enables users to directly access the Oracle® database which the company had chosen as the platform for its data warehouse. During an initial test run, the company analyzed data from customers. The results were impressive: the test clearly highlighted characteristics of the customer group as well as customer demands and requirements. Furthermore, the system was “simple to interpret and easy to understand for each of HEW’s various hierarchical levels,” claims Peter Goos, who works in the company’s marketing and sales management department.

HEW’s data warehouse offered the foundation on which customer classifications could be built. Specific values such as postal codes, towns, streets and house numbers, and electricity rate bracket, were derived from 11,035 pieces of historical data, from a sample of 5,000 customers from the company’s two largest customer groups, plus all of the customers in its smallest customer group. Specially purchased microgeographic characteristics of the customer addresses were also added.

“Before beginning the actual analysis, we used PASW Statistics cross tabulations to establish whether our data were useful—in other words, to find out how much the characteristics chosen tell us about the customer typology,” explains Anke Wilzo, discussing the next stage in the process. Eight indicators proved to be unsuitable, including the characteristic of purchasing power (which is high throughout Hamburg) and region type (which characterizes all Hamburg’s citizens as living in a conurbation).

Results

In the next step, PASW Modeler was used to analyze all of the characteristics simultaneously. The software’s decision trees and profiles produced important findings. While it was still not possible to make clear distinctions for the large customer groups, the distribution of the characteristics in the smallest customer group could be associated with the individual rate types with an accuracy of 90 percent.

More than 183,000 addresses from 60 postal districts were entered into the final model. The result: 3,600 customers will now receive personal offers by post.

“Four more projects are planned in the near future using PASW Modeler,” says Wilzo. The plan is to use historical data to calculate how prepared customers are to change suppliers, thereby estimating potential customer losses. Furthermore, the company aims to derive typical customer groups from customers’ “lifetime value”—the entire sales which a customer generates for the company—and analyze behavioral trends within these groups. Finally, the company wants to find a way to use existing data to assess the risk involved when concluding a contract with a new customer.

Peter Goos is realistic in his assessment of the application:
“Our search for a suitable model has shown that PASW Modeler analyses do not offer HEW complete knowledge. Nevertheless, they provide an important aid in deciding which data provide us with the information we require for CRM purposes, and at which point we have to purchase additional information in order to conduct the relevant assessment.”