Originally appeared in MSI Magazine
May 1, 1999

Yesterday’s News

Must enterprise systems always make money? Just as you’re racing down the homestretch, completing that migration from outdated, poorly-documented, legacy systems that were never Y2K-compliant, somebody finally tells you what you’ve probably known all along, but were too ashamed to admit. Your ERP migration probably won’t pay itself off, according to conventional metrics.

To judge from the tens of billions of dollars invested in new manufacturing enterprise systems over the past decade, a sane-minded business person should be excused for thinking that there must be some cost benefit in all this. Otherwise, it would be unlikely that you would be reading this magazine. And you would have laughed at those high-paid consultants who suggested that your company turn itself upside down, reengineering stovepipe processes and ripping out those heavily-patched legacy systems with code that was no longer decipherable.

Therefore, it was inevitable that somebody would finally study whether these investments delivered positive returns on investment (ROI). According to the Cambridge Information Network, an offshoot of systems integrator Cambridge Technology Partners, you—or your CFO—had better be patient. They found most ERP investments requiring up to four years to pay themselves back. A recent Meta Group study was, on the surface, even less optimistic. Sixty percent of their sample group reported negative ROIs.

Admittedly, Meta’s numbers had a few surprises. Exhibit One: smaller companies pay through the nose on ERM (Meta’s acronym, reflecting the operational focus of most ERP systems). Manufacturers with revenues under $200 million paid, on average, 4.6% of corporate revenues over the life of their enterprise system projects, while larger, billion dollar-plus counterparts expended less than 0.9%. Evidently, this confirms why ERP vendors have been challenged in devising lower-priced implementations for midsized manufacturers.

Yet, these findings were in spite of the fact that large, global enterprises, usually implement broader, deeper functionality, generally relying more on pricey consulting services to implement their projects. According to the survey, SAP customers, which tended to skew larger in the spectrum, averaged spending over half their implementation dollars in consulting services compared to only 22% for Baan, whose customers (outside Boeing) were usually far smaller.

Those results aside, however, Meta’s study confirmed what more astute manufacturers have long considered common sense. Namely, that ERP systems are not implemented for their ROIs. Beyond Y2K issues, manufacturers have embraced ERP solutions because they served as the catalyst to bring enterprises together, providing improved visibility over internal processes, and to varying extents, more reliable snapshots of their supply chains.

Common sense dictates that organizations make better decisions when they have better information. It’s the inverse of the age-old IT maxim, “Garbage in, garbage out.” And better information helps organizations make faster decisions that can lead to faster times-to-market, more satisfied customers, and improved market penetration. Still, the question persists: how do you place a value on that?

The real issue in the Meta findings is the relevance of conventional ROI metrics, period. ROI figures are usually based on labor savings. Yet, the notion of quantifying investments solely by cost savings sounds rather defensive in an era where industry leaders are winning through rapid innovation and conquest of market share .

For veterans of manufacturing, the ROI debate is yesterday’s news. “Like motherhood and apple pie, the benefits of automation are easy to describe but hard to justify… [Direct] labor’s share of the pie has dropped to 10%, while materials and overhead have climbed to 55% and 35% respectively.” Sound familiar? This columnist wrote those words back in 1986, when the debates were centering over the benefits of MRP and CADCAM systems, while “soft” factors such as the untracked costs of poor quality or managing excess inventory buffers to compensate for quality or supply disruptions were generally dismissed by corporate accountants.

Obviously, not all CAD or MRP/ERP projects proved successful. Yet, the horror stories of ERP projects careening out of control did not necessarily mean that “soft” numbers for cost justifying projects were invalid. Instead, the lesson was that poorly managed projects will always generate dismal ROIs, no matter how the numbers have been produced. The bottom line? Keep projects focused on competitive goals, then analyze whether the ROI numbers have much to do with achieving those goals.