Why ROI Is So Elusive

Users spend big bucks on database software and analytical tools because they're just too important to skimp on. Yet once the software is installed, it can be a real challenge to justify the cost. That's because companies get hooked into licensing schemes that don't favor their enterprises, or they buy too much capacity as a buffer against potential peak-time use. Adding to the problem is that analytical applications aren't fully exploited because end users aren't well trained on them - thus limiting the software's value.

"One can roll out an application with great intentions and lofty vision, only to be confronted with the harsh reality that other users just aren't interested," says Mike Tucker, manager of laboratory services at Brooklyn, N.Y.-based energy utility KeySpan Corp.

End users often complain that they don't have time to learn how to operate analytic applications, he says. But given the high cost of such software and the hardware to support it, it can be important to get as many users on board as possible to justify the cost.

KeySpan recently installed analysis applications from Vigilance Inc. in Sunnyvale, Calif., a project that required the utility firm to buy two Windows NT servers.

Cost-conscious companies want to know how an application will produce a return on investment at the outset, but that requires an active commitment from employees to use and exploit the software. "The application will probably not accomplish the savings on its own," says Tucker.

Figuring out whether the database or analytical software is providing payback can be a challenge, says Eric Bloom, vice president of IT at Chadds Ford, Pa.-based Endo Pharmaceuticals Holdings Inc. At his firm, market research personnel create crucial reports for executives, and databases support processes that are vital to the company - even though it's hard to judge how they affect the bottom line, Bloom says.

In general, users also have to navigate through often complex vendor pricing formulas to get the most for their money, says Frank Gillett, an analyst at Cambridge, Mass.-based Forrester Research Inc. "For instance, companies that buy database products by server CPU capacity have to buy with peak usage levels in mind, as opposed to average usage levels," he says. "That means a percentage of the database capacity usually lies untouched."

Capacity vs. Reality

The Web has exacerbated the licensing uncertainty. A company never knows when a flood of queries might come in through a Web portal, raising the risk that a mission-critical e-commerce application could crash. The result: Companies overbuy capacity just to be safe, Gillett says.

Then there are licenses where pricing is based on the number of end-user seats. At one large southern manufacturer that runs an Oracle Corp. database, it seems that there are always fewer actual users than the number of seats purchased, says a senior IT professional at the firm, who requested anonymity. Companies that purchase their licenses by user are always struggling to get the most value from their contracts.

At his shop, most licenses are based on the number of concurrent users, while newer licenses are based on the number of processors, which means that "you have to price things multiple ways to ensure you are paying the least for the particular use of the database," he says.

Oracle's user-based license agreements give customers little flexibility; licenses are issued for a given department, and they can't be shared across the company. Sometimes the manufacturer can sign a network license combining all seats into a single pool and then divvy them up, but this can be cost-prohibitive. On the other hand, when buying based on CPU capacity, a user with a less-powerful PC server may wind up paying more than a user with a mainframe or Unix box.

"The database vendors just have to recognize that their schemes can adversely affect the cost of licensing on certain platforms," says Joe Imbimbo, an Oracle applications database administrator at New York-based recruiting firm TMP Worldwide Inc.

If pricing schemes didn't favor any particular hardware platform, then "everyone will have a level playing field," he says.

Special Report


Taming Data Chaos

Stories in this report:

Copyright © 2002 IDG Communications, Inc.

It’s time to break the ChatGPT habit
Shop Tech Products at Amazon