InformationWeek
Defining The Business Value Of Technology
May 24, 1999
 Print this story |
continued...page 3 of 3
| Related links: |
|
|
| And from our sister publications: |
|
|
It's trickier to quantify the indirect benefits generated by a data warehouse, such as improved customer retention and promotion analysis. In those cases, proving a cause-and-effect link can be difficult. Creating definitions for these benefits--what is a "retained" customer, for example--can help when making pre- and post-data warehouse comparisons, McKnight says.
First Union Corp., a Charlotte, N.C., financial-services company, built its data warehouse two years ago with three initiatives in mind: improving the profitability of existing customers through cross-selling and other activities; retaining valuable customers; and acquiring better, more-profitable customers. First Union then developed annual goals for those initiatives.
"We set targets that were very achievable," says Naras Eechambadi, senior VP of enterprise information management. "And then we asked, `What is the bottom-line impact?'" Because First Union had already calculated the profitability of its customers, figuring out how much the improvements helped in attracting and retaining profit-able customers would yield was possible.
But how was Eechambadi and his team able to calculate just how much the data warehouse would contribute to meeting those goals? "Obviously, you need to make some assumptions or projections," he says of establishing the link between what the data warehouse could provide and how much customer profitability could be improved. "That is really just based on managerial judgment." Data gathered from earlier marketing campaigns provided some empirical evidence of how effective the data warehouse might be, Eechambadi says.
Once the numbers are in place, they should be tied into whatever measurements an enterprise uses, such as internal rate of return, break-even analysis, net present value, and payback period.Georgia-Pacific Corp. in Atlanta, which is installing SAP's Business Information Warehouse as part of a complete R/3 implementation within its paper-manufacturing division, is using equivalent value-added metrics to measure return. But Raden cautions that all such calculations should be "value-based"--that is, understanding that many of the dollar figures are guesses at best.
McKnight even suggests coming up with not one, but three possible return-on-investment calculations for a data warehouse, based on whether the project is successful, partially successful, or fails--with probabilities assigned for each possible scenario. That will help IS managers and company executives judge the riskiness of the potential gain or loss.
And what if the ROI is negative but the IS department is still convinced that the data warehouse system has value? Then IS management will have to make a greater effort to get other areas of the company that might benefit from the data warehouse, such as line of business operations, to sign on to it.
At Horizon Blue Cross Blue Shield of New Jersey, the CIO backed the proposed data warehouse before the health-care company's top management, "because he has seen their value in cost-containment and customer care in similar environments. He knows what they can do for us," Angelillo says. "The track record of an executive sponsor can be more important than ROI numbers."
return to page 1, 2
Photo of Eechambadi by Milton Morris
Back to This Week's Issue
Send Us Your Feedback
Top of the Page
