Ride the wave ­ the tsunami wave of financial transactions, that is. And we've been doing that for years at KeyCorp. Based in Ohio, KeyCorp is one of the largest bank-based financial services companies in the United States. With assets of approximately $81 billion, we provide retail and wholesale banking, investment, financing and money management services to more than 7 million individuals and companies in 46 states. In addition, we have more than 1,000 branches. KeyCorp operates approximately 2,600 ATMs and four telebanking centers which provide financial products and services by phone around the clock, handling more than 4.5 million customer calls in 1997.

In 1996 we had a growing customer base, extensive service diversification and an increasing use of sophisticated market modeling and forecasting tool; so we turned our attention to our business intelligence support systems. In order to study existing customer relationships, we really needed to be able to view the full breadth of our customer information and the accounts that they have at KeyCorp.

Up until that time, information was delivered directly from our operational systems running on older IBM ES/9000 computers. Whenever there was contention over CPU cycles on these computers, the day-to-day banking operations, of course, always took precedence over marketing and customer relationship management applications.

But since our customer base and transactions were growing to tidal wave proportions, we saw the need for a dedicated information system for business intelligence applications. We built a new data warehouse that consolidated customer and account information from each line of business, resulting in a 400 million row, 1.2 terabyte data warehouse. This data warehouse now provides a consistent information base for the marketing analysis and customer relationship management applications used by more than 2,000 employees across KeyCorp.

When we initially decided to build the data warehouse, we tested the IBM S/390 server as well as several UNIX systems. We found that warehouse maintenance was much easier with the S/390 because, for one, the data (which originated from the ES/9000-9021) could remain on the mainframe platform. But more than that, we needed a data warehouse that could also securely and reliably support a number of critical business applications.

In production since January 1997, our data warehouse does just that. For example, an application called KeyLead gives representatives in the branches a view of all customers in their branch and the bank's total relationship with these customers. KeyLead highlights the next set of opportunities that the representatives should be presenting to the customers and provides measurements to determine an increase or decrease in the profitability of the customer relationship.

Another application, Small Business Workstation is used by mobile sales representatives targeting small business accounts. KeyCorp also runs an Automatic Lead Identification program to support sales of cash management services to corporate clients. The bank's Prospecting Tool Kit helps its marketing team select and purchase mailing lists for new prospects.

We also use of a variety of tools for ad hoc query and reporting including IBM Query Management Facility (QMF), IBI FOCUS, SAS and Merant DataDirect Explorer, which are used by the analysts in the bank's various lines of business. Exchange Application's VALEX is used by the marketing organization to manage direct mailing campaigns. We use Hyperion Essbase in corporate finance, and Cognos serves the needs of our client services team who can look at whether customers prefer using an ATM, a branch or a voice response unit. Our experience with the IBM S/390 environment came in handy when deploying the data warehouse.

Because we partitioned the database to take advantage of the multiple engines, users saw improved performance and reliability, in addition to the benefit of having a dedicated query environment. The use of logical partitions (LPARs) allowed KeyCorp to run development, test and production systems concurrently. Recently our team configured the S/390 servers in a parallel sysplex cluster for superior scalability and availability. In the near future, we plan to utilize DB2 data sharing for the data warehouse, which will further improve performance.

One of our primary concerns was that the data warehouse integrate easily with the bank's existing marketing applications. The DB2 data warehouse on the S/390 was able to meet this criteria with its support for standard connectivity protocols such as TCP/IP and ODBC. The benefits of this included a three-second query response time and KeyCorp's lead generation activities. With the data warehouse we have moved from primarily conducting direct mail marketing campaigns to leveraging, from a marketing perspective, most of the channels our customers use.


Troy D. Thomas, Senior Vice President of Customer Warehouse, KeyCorpThe decision we made in 1996 to go with S/390 is one that we haven't regretted. We were able to continue work on our daily tasks. We didn't need to stop everything to move to a new environment. Other companies that have struggled with other technologies have gotten away from building their data warehouse. But we made the decision to go with technology that would positively impact our business and show results. And as a matter of fact, after conducting a ROI study we learned we had realized a 364 percent return on our marketing investment while doubling net income on products sold.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access