Columns

Three little words

• By Tony Baer
• June 30, 2003

Maybe it’s the economy, but in the past few years, the three little words “Return on Investment” have come to haunt technology practitioners. Given the spending excesses of the dot-com era, the new ROI orthodoxy is certainly understandable.

In the business intelligence (BI) world, however, the importance of ROI in cost justifying new projects may be overrated. At a recent conference panel session on the topic, we polled a room full of BI professionals on whether they justified projects using ROI analysis. Surprisingly, few hands went up. As it turned out, we were asking the wrong question.

For starters, ROI cases never tell the full story. Yes, they can pinpoint “hard” costs, such as the cost of acquiring software, computing/networking infrastructure and implementation labor. Additionally, they can be used to compare the costs of competing vendor solutions or the savings possible from productivity improvements. Such hard numbers might be necessary if top management entertains serious doubts about any project.

In some cases, hard numbers might be enough. For instance, if your company is in a market that is driven by cost reduction, pennies saved are pennies earned. Just ask any automotive manufacturer trying to make money by offering models at 0% interest and no money down for five years.

However, in other cases, “easy” numbers won’t even begin to tell the story. For instance, standardizing financial metrics in a decentralized organization might save time and effort during the reconciliation process. But those gains could easily pale if the reports also uncover opportunities to eliminate unnecessary discounting practices that are costing profits.

Additionally, some systems won’t deliver positive ROI; instead, they become part of the cost of doing business. That was the conclusion of a 1999 Meta Group study that evaluated ERP projects. According to the study, most companies needed modernized back-office systems to remain competitive.

Rather than asking whether ROI analyses are necessary to justify new BI projects, the more appropriate question to ask is whether such analyses can keep projects on track. In essence, ROI analyses might be perfect for providing decision support about decision support, gauging how effective they are at uncovering new business opportunities, and satisfying enterprise goals or end-user needs.

For instance, ROI analyses could help to determine whether reports or queries answer the right questions. ROI metrics could be derived from factors such as the rate or value of use. Arbitrary values might be assigned to utilization rates or actual values could be attributed to new opportunities for reducing costs or increasing revenue.

An extreme example would be a $1 million BI system that is used exactly once. Despite appearing to be a poor investment at first glance, it might have spawned a query that revealed an undiscovered $2 million savings or a $5 million market opportunity. Conversely, a frequently used $75,000 system that looked like a bargain might be judged a waste if it produced answers that were already known.

Consequently, while ROI analyses of BI systems rarely generate the hard figures that stand up to standard accounting principles, they could provide insight into the value of the investment.

What’s even more difficult -- and more significant -- may be the “soft” benefits that are not easily quantified. For instance, what is the real value of a system where resulting productivity savings enable professionals to spend more “quality time” with their customers? If, for instance, account representatives are freed up to spend an additional 20% of their time with their clients, resulting in a 15% increase in sales or customer retention rates, could all those improvements be attributed to the added attention, or to improvements in factors such as products or promotions?

Ironically, this is where “hard” data may come to the rescue. If the system produces analyses that more effectively identify opportunities to improve product mix or pricing levels, the ROI becomes far more believable.

In retrospect, maybe it shouldn’t have been surprising that BI professionals at our session performed few ROI analyses. Like most technology professionals, they are busy and have little time to perform the rigorous cost/benefit studies that the experts are calling for. Nonetheless, that does not discount the value of those three little words in gauging the effectiveness of BI projects after they have gone live.