InformationWeek.com

Welcome Guest. | Log In| Register | Membership Benefits

May 24, 1999

Print this story

Warehouse ROI

Data warehouses are getting the same scrutiny as other projects

By Rick Whiting

Related links:

Maximize Software ROI

Software Resource Center

And from our sister publications:

TechWeb What's The Investment Worth?

InternetWeek Measuring ROI For The Top Line Of The Business

s it possible to calculate the financial return on investment of a data warehouse? Perhaps more important, should companies even try?

Since data warehouses first appeared in the early 1990s, they've largely been viewed as exempt from the tough return-on-investment or payback analysis to which other IT acquisitions are subjected. As late as 1997, a Forrester Research survey found that only 16 of 50 companies were conducting any kind of dollar-based ROI measurements before implementing a data warehouse.

The experience at CH2M Hill Inc., a Denver environmental engineering company, is typical. "We built a small data warehouse as a research project, and it just grew into the real thing," says data warehousing project manager Scott Crownover. But did Crownover and his crew do any kind of detailed, preassembly ROI exercise? "We sure didn't," he says.

That may be changing. "Data warehouses are no longer being justified with mantras, ambiguous concepts, and qualitative factors," says William McKnight, managing principal at McKnight Associates, a consulting firm. He says the trend is toward applying some of the same ROI and payback analysis that governs most capital expenditures.

Data warehouses provide benefits such as better inventory management, more effectively targeted sales and marketing efforts, and improved customer satisfaction. The problem is trying to attach a dollar figure to such gains.

"Data warehousing is a second- or third-level benefit--that is, it leads to other processes that in turn result in reduced costs and higher sales," McKnight says. A company might use a data warehouse to identify new sales prospects or to uncover hidden costs at a manufacturing plant. But those potential gains will never be realized unless a sales representative calls on the prospects or plant managers take steps to eliminate the deficiencies.

That's the situation at Horizon Blue Cross Blue Shield of New Jersey in Newark, which is more than a year into building a data warehouse with membership, claims, and health-care provider information. When the system is in full production, Horizon employees will be able to tap into its store of data to analyze the performance of health-care providers--spotting when a doctor is repeatedly prescribing high-priced drugs, for example, or ordering too many expensive tests.

Time For Action
But for the data warehouse to bring returns to a company, someone has to act on its findings, such as reining in the doctor with the fondness for top-of-the-line prescriptions, or implement whatever best practices the data warehouse uncovers. "Quantifying the result of someone's acting upon those findings is difficult," says Richard Angelillo, Horizon's data warehouse project manager. That inability to identify a hard-and-fast payback is why Angelillo's team didn't attempt to calculate the project's ROI. "I'm not a big believer in the validity of that type of exercise," he says.

However, the decision about whether ROI numbers are needed rarely rests in the hands of IT management. CEOs, CFOs, and other top managers sometimes demand proof of a financial return for major IT projects. In some companies, "doing it by the numbers" is ingrained in the culture, particularly in the finance and retail industries.

"There are some companies that drive everything by ROI," says Neil Raden, president of Archer Decision Sciences, a data warehouse and decision-support-system consulting firm. High-profile cases of over-budget or failed data warehouse and enterprise resource planning implementations have also made top management leery of big-ticket IT projects.