Columns

Easy money?

• By Tony Baer
• September 1, 2002

The rise in housing prices is the latest proof that nature abhors a vacuum. No matter how dire the circumstances, there's always money floating around for something. Yesterday, the money went into stocks; today, it's housing.

The same is true in IT organizations. Yesterday, ''strategic investments,'' like ERP, CRM, e-business transformation, knowledge management or zero latency were often green-lighted. Today, the conventional wisdom is that software projects must demonstrate tangible ROI. But how true is that?

A technology tale from 20 years ago provides some clues. During the early 1980s, the economy was in a recession and U.S. manufacturers were taking a beating from the Japanese. At a Detroit automotive industry conference, a computer graphics program drew a flurry of attention. The program was a factory simulation system showing the impact of various factors on an assembly line. As one of the earliest forms of decision support tools, factory simulation software came of age during that recession because manufacturers wanted to hedge their multimillion-dollar buying decisions.

Consequently, in this recession, decision support remains one of the few IT investments consistently making the cut. Business intelligence (BI), a form of decision support, remains viable because its goals are understandable: Make businesses more effective by using the information already at their disposal. Assuming the costs of the project are reliably estimated, the technology challenges are documented, the goals well defined and the scope well bounded, about the only objection that could be raised is how to ensure such systems remain relevant to the organization. It's no wonder that BI firms like Business Objects and Hyperion remain in the black -- a fairly rare accomplishment among software vendors today.

Using similar criteria, portal initiatives should pass muster because they too are aimed at making better use of existing resources. Like BI systems, they can help firms to leverage the millions of dollars they have invested in SAP, PeopleSoft, Siebel or similar projects.

Yet related initiatives to portals, such as content or knowledge management, haven't fared quite so well. These initiatives all too often degenerate into enterprise information scavenger hunts and organizational reengineering. Conversely, portal projects can be confined to developing useful front ends with ''lite'' content management features, classifying information that is already there. More importantly, when portals work, users and top management can readily view the results on their screens. And from an ROI standpoint, portals can be very easily linked to ROI improvements such as increased sales or decreased response times.

In other cases, strict ROI criteria may not always be the deciding factor. For instance, the ROI for development tools like software testing are quite tangible, because it is relatively straightforward to document productivity gains when QA processes are automated. But with some gains, like decreased software bugs, it may be hard to assign a firm dollar value.

The picture grows hazier when you look at other development tools. For instance, IDEs with the right tools, languages and visual navigation should be more productive. However, unless your organization is willing to spin its wheels doing blind taste tests, the savings will be impossible to prove. The situation is similar with other development tools, such as analysis and design. Although results such as code reuse should demonstrate an ROI, the degree of reuse typically depends more on human nature than on the effectiveness of the tools.

What about ''Big Box'' initiatives? For example, server consolidation, enterprise software harmonization and EAI should all sound appealing to the bottom line. While each of these measures should produce real savings, they will be tempered by cultural and organizational factors that could make the transition bumpy.

Consequently, when money is tight, it makes sense to use ROI as the first step in ferreting out sure losers. But don't get carried away with numbers, because strict ROI analyses will only tell part of the story.